TECHNICAL REPORT DOCUMENTATION PAGE - STRIDE

TECHNICAL REPORT DOCUMENTATION PAGE

1. Report No.

Project F

2. Government Accession No. 3. Recipient’s Catalog No.

4. Title and Subtitle

Integrated Implementation of Innovation Intersection Designs

5. Report Date

May 2020

6. Performing Organization Code

7. Author(s)

Dr. Nagui M. Rouphail, Ph. D., North Carolina State University

Mr. Chris Cunningham, P.E., North Carolina State University

Dr. William J. Davis, Ph.D., The Citadel

Ms. Shannon Warchol, M.S., North Carolina State University

Ishtiak Ahmed, Ph.D., North Carolina State University

8. Performing Organization Report No.

Project F

9. Performing Organization Name and Address

North Carolina State University

2501 Stinson Drive

Raleigh, NC 27695

The Citadel

171 Moultrie Street

Charleston, SC 29409

10. Work Unit No.

11. Contract or Grant No.

Funding Agreement Number

69A355174710

12. Sponsoring Agency Name and Address

University of Florida Transportation Institute

Southeastern Transportation Research, Innovation, Development and Education

Center

365 Weil Hall,

P.O. Box 116580

Gainesville, FL 32611

U.S Department of Transportation/Office of Research, Development & Tech

1200 New Jersey Avenue, SE

Washington, DC 20590

United States

13. Type of Report and Period Covered

5/16/2017 to 5/21/2020

14. Sponsoring Agency Code

15. Supplementary Notes

16. Abstract

This study evaluated the performance of pedestrian-bicycle crossing alternatives at Continuous Flow Intersections (CFI). Further, a

comparison was also performed of CFI crossing types against a standard intersection designed to provide an equivalent volume-to-

capacity ratio. Three CFI crossing alternatives were tested, namely Traditional, Offset, and Midblock crossings. In total, 24

alternative scenarios were generated by incorporating two bicycle path types, two right-turn control types, and two CFI types.

These scenarios were analyzed through microsimulation on the basis of stopped delay, travel time, and number of stops.

Simulation results revealed the Offset crossing alternative incurred the least stopped delay for all user types. The Traditional

crossing generated the least number of stops. The Midblock crossing can be considered as a supplement to either the Offset or

Traditional crossing depending on the specific origin-destination patterns at the intersection. The exclusive bicycle path performed

better than the shared-use path in most cases. However, these general trends may vary significantly at the route-level analysis.

When compared with an equivalent standard intersection, aggregated results showed significant improvement for all CFI crossing

types with respect to travel time and stopped delay, but the standard intersection had an equal or fewer number of stops. Future

research includes incorporating pedestrian-bicycle safety, comfort, and the relative effects of these crossing alternatives on

vehicular operations. Finally, a summary of the multimodal aspects of Reduced Crossings IU-Turn (RCUT) intersections, as

studied by researchers at The Citadel is provided in Appendix B of this report. The reduction of the number of conflict points to as

few as eight provided by RCUT intersection configurations was found to be beneficial to all roadway users including, motor

vehicle traffic, pedestrians, bicyclists, and transit users.

17. Key Words

Continuous Flow intersection, Displaced left turn, Alternative

intersection, Pedestrian, Bicycle, RCUT Multimodal effects

18. Distribution Statement

No restrictions

19. Security Classif. (of this report)

20. Security Classif. (of this page)

21. No. of Pages

78 Pages

22. Price

Form DOT F 1700.7 (8-72) Reproduction of completed page authorized

Integrated Implementation of Innovative Intersection Designs

ii

DISCLAIMER The contents of this report reflect the views of the authors, who are responsible for the facts and

the accuracy of the information presented herein. This document is disseminated in the interest

of information exchange. The report is funded, partially or entirely, by a grant from the U.S.

Department of Transportation’s University Transportation Centers Program. However, the U.S.

Government assumes no liability for the contents or use thereof.

ACKNOWLEDGEMENT OF SPONSORSHIP AND STAKEHOLDERS This work was sponsored by a contract from the Southeastern Transportation Research,

Innovation, Development and Education Center (STRIDE), a Regional University Transportation

Center sponsored by a grant from the U.S. Department of Transportation's University

Transportation Centers Program.

Funding Agreement Number - 69A3551747104


iii

LIST OF AUTHORS

Lead PI:

Dr. Nagui M. Rouphail, Ph.D.

Distinguished University Professor

Department of Civil, Construction, and Environmental Engineering,

North Carolina State University

[email protected]

ORCID Number: 0000-0002-2420-9517

Co-PI:

Chris Cunningham, PE

Director, Highway Systems Group, ITRE

[email protected]

Additional Researchers:

William J. Davis, Ph.D., PE

Distinguished University Professor

D. Graham Copeland Professor of Civil Engineering

The Citadel University

[email protected]

ORCID Number: 0000-0002-3812-8654

Shannon Warchol, PE

Senior Research Advisor

Highway Systems Group, ITRE

[email protected]

ORCID Number: 0000-0003-2083-5868

Ishtiak Ahmed, M.S.

Graduate Research Assistant

CCEE Department, NC State University

[email protected]

mailto:[email protected]






iv

TABLE OF CONTENTS DISCLAIMER .................................................................................................................................. ii

ACKNOWLEDGEMENT OF SPONSORSHIP AND STAKEHOLDERS .................................................... ii

LIST OF AUTHORS ......................................................................................................................... iii

TABLE OF CONTENTS ................................................................................................................... iv

LIST OF FIGURES ........................................................................................................................... vi

LIST OF TABLES ............................................................................................................................ vii

ABSTRACT .................................................................................................................................. viii

EXECUTIVE SUMMARY ................................................................................................................. ix

1. INTRODUCTION ................................................................................................................... 10

1.1 Background ....................................................................................................................... 10

1.2 Research Objectives .......................................................................................................... 11

1.3 Scope ................................................................................................................................. 12

1.4 Organization of the Report ................................................................................................ 12

2. LITERATURE REVIEW ........................................................................................................... 13

2.1 Application of Microsimulation in Pedestrian-Bicycle Studies ........................................... 13

2.2 Simulation of Pedestrian-Bicycle Crossing Alternatives at CFIs ......................................... 13

2.3 Frameworks for Analyzing Pedestrian-Bicycle Crossings at CFIs ........................................ 15

2.4 Qualitative Assessment of Pedestrian-Bicycle Crossings at CFIs ........................................ 16

2.5 Performance Measures ..................................................................................................... 18

2.6 Summary ........................................................................................................................... 19

3. RESEARCH METHODOLOGY ................................................................................................. 20

3.1 Crossing Alternatives ......................................................................................................... 20

3.1.1 Traditional Crossing .................................................................................................... 20

3.1.2 Offset Crossing ........................................................................................................... 21

3.1.3 Midblock Crossing ...................................................................................................... 22

3.2 Analysis Using Microsimulation ......................................................................................... 23

3.2.1 Base CFI Geometry Model .......................................................................................... 23

3.2.2 Adjacent Intersection Geometry ................................................................................ 24

3.2.3 Pedestrian and Bicycle Model Construction ............................................................... 25

3.2.4 Scenario Generation ................................................................................................... 26


v

3.2.5 Traffic Volume ............................................................................................................ 30

3.2.6 Signal Timing .............................................................................................................. 30

3.3 Equivalent Standard Intersection Modeling ...................................................................... 32

3.4 Comparison of Crossing Alternatives................................................................................. 32

4. RESULTS .............................................................................................................................. 34

4.1 Route Level Comparison of Crossing Alternatives ............................................................. 34

4.1.1 Stopped Delay ............................................................................................................ 34

4.1.2 Travel Time ................................................................................................................. 36

4.1.3 Number of Stops ........................................................................................................ 37

4.1.4 Deviations from Common Trends ............................................................................... 38

4.2 Aggregated Results ............................................................................................................ 40

4.2.1 Stopped Delay ............................................................................................................ 40

4.2.2 Travel Time ................................................................................................................. 43

4.2.3 Number of Stops ........................................................................................................ 45

4.3 Summary ........................................................................................................................... 47

5. CONCLUSIONS ..................................................................................................................... 48

5.1 Summary of the Simulation results ................................................................................... 48

5.2 Recommendations ............................................................................................................ 49

6 REFERENCE LIST ........................................................................................................................ 51

7 APPENDICES ............................................................................................................................. 53

7.1 APPENDIX A – Performance Measures by Route Type and Alternative Scenarios ......... 53

7.2 Appendix B – Pedestrian, Bicycle and Transit Considerations for RCUT Intersections ... 62

7.3 Appendix C – Supplementary Diagrams of Pedestrian Crossings at Innovative Geometric

Designs .................................................................................................................................... 72

7.4 Appendix D – Summary of Accomplishments .................................................................... 78


vi

LIST OF FIGURES Figure 1-1: Schematic of a CFI or DLT intersection ...................................................................... 11

Figure 2-1: CFI Geometry and phase sequence for (a) Traditional crossing (b) Offset crossing

(Courtesy, Coates et al., 2014) .................................................................................................... 15

FIGURE 2-2: Modified Offset crossing (Chlewicki, 2017) ............................................................. 17

Figure 2-3: Bicycle crossing options at a CFI (Courtesy: UDOT, 2013) ......................................... 18

Figure 3-1: (a) Schematic of Traditional crossing (b) An example of this crossing type in West

Valley City, UT ............................................................................................................................. 21

Figure 3-2: Schematic of Offset crossing (b) An example of this crossing type in Loveland, CO .. 22

Figure 3-3: Schematic of a Midblock crossing ............................................................................. 23

Figure 3-4: Base geometry model of CFI ..................................................................................... 24

Figure 3-5: Configuration of the adjacent intersections .............................................................. 25

Figure 3-6: Histograms of (a) walker (b) jogger and (c) bicycle speeds ....................................... 26

Figure 3-7: Control types for right turning vehicles (a) Channelized yield control (b) Channelized

signal control ............................................................................................................................... 28

Figure 3-8: Ring barrier diagram and movements in a partial CFI ............................................... 31

Figure 3-9: Quadrants and origin-destination points of pedestrians and bicycles (shared) ........ 33

Figure 4-1: Common trend of stopped delay across different route types and CFI models ........ 35

Figure 4-2: Common trend of travel time across different route types and CFI models ............. 37

Figure 4-3: Common trend of number of stops across different route types and CFI models .... 38

Figure 4-4: Deviations from the common trend in terms of (a) travel time (b) stopped delay and

(c–d) number of stops ................................................................................................................. 40

Figure: 4-5 Stopped Delay aggregated over all routes (a) Partial CFI (b) Full CFI ......................... 41

Figure 4-6 Travel time aggregated over all routes (a) Partial CFI (b) Full CFI ............................... 43

Figure 4-7 Number of Stops aggregated over all routes (a) Partial CFI (b) Full CFI ...................... 45


vii

LIST OF TABLES Table 2-1: Performance measures used in past studies .............................................................. 19

Table 3-1: VISSIM Model Combinations Based on Variable Inputs .............................................. 27

Table 3-2: Right turn control types at different CFIs ................................................................... 29

Table 3-3: Vehicular volume data used in the CFI models ........................................................... 30

Table 3-4: category of pedestrian and bicycle (shared) routes ................................................... 33

Table 4-1 Post Hoc Tukey Test Results on Difference of Stopped Delay between Different Pair of

CFI Crossing Alternatives ............................................................................................................. 42

TABLE 4-2 Post Hoc Tukey Test Results on Difference of Travel Time between Different Pair of

CFI Crossing Alternatives ............................................................................................................. 44

TABLE 4-3 Post Hoc Tukey Test Results on Difference in Number of stops between Different Pair

of CFI Crossing Alternatives ......................................................................................................... 46


viii

ABSTRACT This study evaluated the performance of pedestrian-bicycle crossing alternatives at Continuous

Flow Intersections (CFI). Further, a comparison was also performed of CFI crossing types against

a standard intersection designed to provide an equivalent volume-to-capacity ratio. Three CFI

crossing alternatives were tested, namely Traditional, Offset, and Midblock crossings. In total,

24 alternative scenarios were generated by incorporating two bicycle path types, two right-turn

control types, and two CFI types. These scenarios were analyzed through microsimulation on

the basis of stopped delay, travel time, and number of stops.

Simulation results revealed the Offset crossing alternative incurred the least stopped delay for

all user types. The Traditional crossing generated the least number of stops. The Midblock

crossing can be considered as a supplement to either the Offset or Traditional crossing

depending on the specific origin-destination patterns at the intersection. The exclusive bicycle

path performed better than the shared-use path in most cases. However, these general trends

may vary significantly at the route-level analysis. When compared with an equivalent standard

intersection, aggregated results showed significant improvement for all CFI crossing types with

respect to travel time and stopped delay, but the standard intersection had an equal or fewer

number of stops. Future research includes incorporating pedestrian-bicycle safety, comfort, and

the relative effects of these crossing alternatives on vehicular operations.

Finally, a summary of the multimodal aspects of Reduced Crossings IU-Turn (RCUT)

intersections, as studied by researchers at The Citadel is provided in Appendix B of this report.

The reduction of the number of conflict points to as few as eight provided by RCUT intersection

configurations was found to be beneficial to all roadway users including, motor vehicle traffic,

pedestrians, bicyclists, and transit users.

Keywords: Continuous Flow intersection, Displaced left turn, Alternative intersection,

Pedestrian, Bicycle, RCUT Multimodal effects


ix

EXECUTIVE SUMMARY

Continuous Flow Intersections (CFIs), also known as Displaced Left-turn intersections (DLTs) have

grown popularity in many countries primarily due to its reduced number of signal phases for

vehicles as it removes the conflict between left-turn and opposing thru movements. However, due

to its large footprint and unconventional displaced left-turn movement, pedestrians and bicycles

experience unique challenges at this type of intersection. Therefore, it is important to evaluate

different pedestrian-bicycle crossing alternatives at CFIs. The main objectives of this research are

two-fold. The first objective is to evaluate the mobility performance of pedestrian and bicycle

crossing facilities at CFIs. In addition to crossing alternatives, different alternative options for bicycle

path type, vehicle right-turn control type, and CFI type are considered to generate various

alternative scenarios. The second objective is to compare the CFI crossing alternatives with a

standard intersection crossing of equivalent volume-to-capacity ratio. Three types of CFI crossing

alternatives – Traditional, Offset, and Midblock crossing – are tested in this study. In total, 24

alternative scenarios were generated by varying the CFI type, right-turn control type, and the

bicycle path type. All these alternative scenarios were analyzed through microsimulation based on

average stopped delay, travel time, and number of stops.

Results from the simulation runs revealed that for most route types, Midblock crossing generated

the highest stopped delay, while the Offset crossing generated the lowest among the three crossing

alternatives. Travel time showed the same trend as the stopped delay for most route types. Offset

crossing generated the highest number of stops compared to the other two crossing types primarily

due to the increased number of stages. Traditional crossing exhibited the least number of stops due

to its straightforward configuration. Overall, exclusive bicycles generated significantly lower

stopped delay, travel time, and number of stops than bicycles on the shared-use path. Route level

analysis revealed several deviations from these common trends for some route types.

Post-Hoc Tukey test revealed the difference in stopped delay incurred by Midblock and Offset

crossing is the most prominent. Except for exclusive bicycle stopped delay and number of stops, the

difference in performance measures for other users between different pairs of crossing alternatives

are statistically significant. When comparing an equivalent standard intersection crossing with a CFI

crossing, a CFI with Traditional or Offset crossing incurred less stopped delay and travel time

because of the reduced number of phases. However, a CFI with an Offset or a Midblock crosswalk

generated a higher number of stops than a standard intersection because of the increased number

of stages. For future research, we recommend testing other innovative crossing options, for

instance, a combination of Midblock and Traditional crossing at a CFI. Further, it is suggested to

investigate pedestrian-bicycle safety, comfort, and the relative effects of these crossing alternatives

on vehicular operations.

A summary of the multimodal aspects of RCUT intersections was also investigated by The Citadel

(see Appendix B). Results indicated that a reduction of the number of conflict points to as few as

eight provided by RCUT configurations was found to be beneficial to all roadway users including,

motor vehicle traffic, pedestrians, bicyclists, and transit users.


10

1. INTRODUCTION

1.1 Background Addressing pedestrian and bicycle mobility in conjunction with traffic operation and

motor vehicle efficiency is a critical concern in the design and configuration of

transportation infrastructure. Pedestrians and bicycles often experience excessive

delays, have higher risks, and safety concerns at signalized intersections due to the

concentrated number of conflict points with motorized traffic flows and other factors

related to cycle length and levels of congestion (Pulugurtha and Imran, 2015). Further,

the overall quality of service for pedestrians and bicycles can be worse in the case of

alternative intersections, which are designed primarily to improve the efficiency of

vehicular movements.

Alternative intersections and interchanges like Continuous Flow Intersections (CFI),

Restricted Crossing U-turns (RCUT), and Diverging Diamond Interchange (DDI) are

gaining popularity all over the world (Jagannathan and Bared, 2005; Warchol et al.,

2017; Holzem et al., 2015). An RCUT reroutes all minor street through and left

movements to turn right at the main intersection, then utilize a U-turn crossover, and

then return to the main intersection to finish the movement. CFI or Displaced Left-turn

Intersection (DLT) is another alternative intersection which is mostly built in urban setup

(Hughes et al., 2010) and at junctions where left-turn demand is expected to be high. In

this system, additional signals upstream of the main intersection are introduced to

move the left-turn lanes to the left of the exit lanes upstream of the main intersection.

Thus, the conflict between the left-turn vehicles and the opposite through vehicles at

the main intersection is transferred to the upstream intersections. Therefore, the main

intersection can run as few as in two phases, which consequently improves the

efficiency of vehicular movements. The signal timing of the upstream intersections can

be designed in such a way that the upstream intersections do not incur additional delays

to the opposing thru movements. A simplified line diagram of a CFI with displaced left-

turn on two legs (Partial CFI) is shown in Figure 1-1.


11

FIGURE 1-1: SCHEMATIC OF A CFI OR DLT INTERSECTION

Despite being favorable to vehicular movements, CFIs pose unique challenges to

pedestrians and bicyclists. The footprint of a CFI is designed larger than an equivalent

standard intersection in order to provide a sufficient radius of curvature for the

displaced left-turn lanes. Further, the crosswalks need to be multi-staged when the

signal timing plan of a CFI has two phases (Coates et al., 2014). On the other hand, while

RCUT intersections reduce conflict points for pedestrians and bicycles, the currently

used crosswalks are multi-staged, and the designs are foreign to most users. Since CFIs

can be built in urban and suburban areas (Hughes et al., 2010) where pedestrian and

bicycle activities are common, it is imperative to investigate different pedestrian and

bicycle crossing options at these alternative intersections. The major focus of this study

is on pedestrian-bicycle mobility at a CFI. It also focuses on pedestrian-bicycle

considerations at RCUT intersections to a limited extent.

1.2 Research Objectives The main objectives of this research are two-fold. The first objective is to evaluate the

mobility performance of pedestrian and bicycle crossing facilities at CFIs. In addition to

crossing alternatives, different alternative options for bicycle path type, vehicle right-

turn control type, and CFI type are considered to generate various alternative scenarios.

The second objective is to compare the CFI crossing alternatives with a standard

intersection crossing designed to provide an equivalent volume-to-capacity ratio.

VISSIM microsimulation tool was used to evaluate all crossing alternatives.

In addition to pedestrian-bicycle crossing alternative analysis at CFIs, a secondary

objective of this study is to review the state-of-the-art practice of pedestrian-bicycle

safety analysis and transit consideration at RCUT intersections. That discussion is

provided in Appendix B of this report.


12

1.3 Scope The scope of this research was limited to pedestrian and bicycle activities at CFIs.

Specifically, the research sought to recommend crossing alternatives for pedestrians

and bicycles at CFIs in terms of different performance measures and also to portray the

contrast with the crossing at equivalent standard intersections. The effect of these

crossing options on vehicular movements is not investigated in this study.

1.4 Organization of the Report This report includes five chapters and two appendices, beginning with this introductory

chapter. Chapters 2 covers the literature review. Chapter 3 presents the research

methodology, including the analysis methods considered and details about the

microsimulation models of alternative scenarios that were generated. Chapter 4 covers

the results from the simulation runs and their interpretations. Conclusions and

recommendations are provided in Chapter 5. There are four appendices that provide

details on performance measures (Appendix A): Safety aspects of RCUT intersections

(Appendix B), supplementary diagrams of pedestrian crossing facilities at innovative

geometric designs (Appendix C), and a Summary of Accomplishments (Appendix D).


13

2. LITERATURE REVIEW

A review of the published literature on pedestrian and bicycle crossing facilities at alternative

intersections (with the major focus on Continuous Flow Intersections or CFIs) is presented in

this chapter. Studies that applied microsimulation techniques for evaluating pedestrian and

bicycle crossing facilities are emphasized. The outcomes are expected to assist in simulating and

evaluating pedestrian crossing facilities at CFIs.

2.1 Application of Microsimulation in Pedestrian-Bicycle Studies The application of microsimulation tools to assess non-motorized users’ safety and

mobility at various types of intersections has been found in several studies. The concept

of simulating pedestrians as vehicles in VISSIM was introduced by Ishaque and Nolan

(Ishaque and Nolan, 2009). It demonstrated the details of pedestrian and vehicular

traffic simulation in a VISSIM environment for a series of standard intersections. Field

data on various parameters such as pedestrian speed, flow rate, compliance, vehicle

travel time, and vehicle flow rate were collected for a route in London, UK. The route

consisted of six intersections in series and a total of nine pedestrian crosswalks. There

was no interaction among the pedestrians so that their performance measures were not

affected unaffected by their density. Calibration and validation of vehicular flow data

were promising and evidenced a successful simulation. However, the simulated speed of

pedestrians had a poor correlation with the observed speed data. Therefore, the model

was recalibrated using a different model, which exhibited significant improvement of

the correlation. Overall, this study described the important elements regarding

pedestrian crossing at a signalized intersection. Holzem et al. (Holzem et al., 2015) also

modeled pedestrians and bicycles in VISSIM as vehicles without any interaction to test

different crossing alternatives at superstreets. Four types of pedestrian crossings are

recognized in superstreets: diagonal, median, midblock, and 2-stage Barnes Dance.

Three types of bicycle crossing facilities, namely direct crossing, vehicle U-turn, bicycle

U-turn – are evaluated as well. Edara et al. (Edara et al., 2015) simulated both Double

Crossover Intersection (DXI) and Diverging Diamond Interchange (DDI) in a VISSIM

environment and the performance measures of pedestrian-bicycles were contrasted

against those for a traditional intersection. Several scenarios for each intersection

geometry were generated by varying traffic volume.

2.2 Simulation of Pedestrian-Bicycle Crossing Alternatives at CFIs The application of microsimulation in assessing CFI performance was found as early as

2005 (Jagannathan et al., 2005). That study demonstrated the assessment of one

pedestrian crossing type – a multi-staged crossing with refuge islands – at three

different geometric designs of the CFI using VISSIM. The geometries modeled were: a

four-legged CFI with two displaced left-turn legs (Partial CFI), a four-legged CFI with four

displaced left-turn legs (Full CFI), and a three-legged CFI (T-intersection) with displaced


14

left turns only at the major street approach. Note that the most common CFI geometry

built in the U.S. is the four-legged Partial CFI.

Linear programming along with microsimulation was utilized in a few studies as well.

Coates et al. (Coates et al., 2014) compared pedestrian safety and mobility for two types

of crossing facilities in a four-legged CFI, namely Traditional and Offset, using VISSIM.

Figure 2-1 (a) and (b) show the geometry of these two crossing facilities along with their

phase sequences. The traditional crossing generated less crossing time than the offset

crossing due to the straightforward structure but incurred additional delays to the

vehicular movements. To tackle this issue, this study proposed an adaptive control

strategy for reducing vehicular delay in a CFI with Traditional crossing. A multi-objective

mixed-integer programming model was proposed by (Zhao et al., 2015) to achieve the

best operational performance of a CFI by changing the CFI type, configuration of the

right-turn lane, distance to the displaced left-turn junction, and signal timing plan.

However, it did not focus on the crosswalk geometries of the CFI. Zhao et al. (2019)

proposed to improve the operation of a CFI by shifting the crossing location of left-

turning bicycles to the midblock location, so there is no conflict with thru traffic. A linear

programming tool was used to optimize the geometry and signal timing, and it was

tested by simulating a real intersection in VISSIM. Constraints added to the linear

programming were based on the phase plan, cycle length, minimum green time, queue

storage capacity for left-turn bicycles and vehicles, saturation flow rate, and degree of

saturation.


15

(a)

(b)

FIGURE 2-1: CFI GEOMETRY AND PHASE SEQUENCE FOR (a) TRADITIONAL CROSSING (b) OFFSET CROSSING

(COURTESY, COATES ET AL., 2014)

2.3 Frameworks for Analyzing Pedestrian-Bicycle Crossings at CFIs Several studies proposed analytical frameworks to evaluate the operation of CFIs in

terms of pedestrian-bicycle crossings and vehicle movements. Wang et al. (2019)

developed an analytical model to calculate pedestrian delay at a CFI for three types of

crossings and tested its accuracy with VISSIM. In addition to a traditional and an offset

crossing, it demonstrated the application of an exclusive pedestrian phase. Only

pedestrian movements on all directions (e.g., thru and diagonals) run in this exclusive

pedestrian phase. However, this study did not consider the delay to vehicles accrued by

pedestrians to cross such a large intersection diagonally. FHWA published an analytical

tool called “Cap-X” (Lochrane, 2011) that compares the performance of eight types of

intersections, including CFIs for different vehicle demands and lane configurations.

Virginia Department of Transportation (VDOT) developed a tool (VJuST, 2017) to analyze

the performance of 26 Alternative Intersections and Interchanges in terms of vehicular

congestion, safety, and pedestrian accommodation for screening purposes. Although

these studies provide a quick sketch-level assessment of alternative intersections, the


16

methods are deterministic in nature and cannot capture the stochasticity involved with

the demand and capacity of the intersections

2.4 Qualitative Assessment of Pedestrian-Bicycle Crossings at CFIs A few studies discussed the pedestrian-bicycle accommodations of the CFI using only

qualitative assessment. Chlewicki (2017) demonstrated the current practice of

pedestrian-bicycle crossing facilities at innovative intersections qualitatively and

proposed several improvement options. In this paper, the term “Innovative Geometric

Designs” (IGDs) was used instead of an alternative intersection. This study introduced

several general principles of a pedestrian crossing at various IGDs, including pedestrian

sight distance, crossing distance, and pedestrian phasing. Exclusive pedestrian phasing

was recommended during the pedestrian-bicycle peak hour if that does not conflict with

vehicular peak hours.

The study discussed pedestrian crossing facilities at seven types of IGDs. Additional

diagrams explaining this study by Chlewicki (2017) is provided in Appendix C. The

summary of the discussion for CFI crossing facilities is provided below:

1. Crossing option at the main intersection: Pedestrian-bicycle crossing at the min

intersection of a CFI can be either multi-staged (Offset) or single staged

(Traditional). This study preferred a multi-staged crossing over a long, single-

staged one since it eliminates the conflicts between turning vehicles and

pedestrians.

2. Midblock Crossing with Median Sidewalk: It proposed a new crossing design

where one stage of crossing occurs upstream from the main intersection and the

other stages occur midblock near the displaced left-turn intersection. According

to this design, the signal at the displaced left turn and the main intersection can

be easily synchronized so that pedestrian-bicycles experience progression

through the two intersections.

This study also proposed modified versions of the crossing alternatives mentioned

above. For instance, it proposed a crossing alternative similar to the Offset crossing,

where one stage to cross a street takes place at the main intersection and another one

at the midblock intersection (Figure 2-2). It also proposed a “Median sidewalk” and

“Reduce median” crossing options. However, pedestrian-bicycles need to cover a long

distance over a narrow median with vehicle movements on both sides according to

these modified designs. Thus, their level of comfort is likely to be degraded.


17

FIGURE 2-2: MODIFIED OFFSET CROSSING (CHLEWICKI, 2017)

Several guidance reports have been published regarding the mobility and safety of

pedestrian-bicycles in a CFI. Among these, the Utah Department of Transportation

(UDOT) and FHWA (UDOT, 2018; FHWA, 2014) discussed the configuration of

pedestrian-bicycle crossing facilities at a CFI along with its signal timing plan. The

objective of the UDOT guidance report was to accelerate the acceptance of CFI use in

Utah. It discussed the configuration of pedestrian-bicycle crossing facilities and

associated signal timing plans. The critical points of the guideline are described below:

• The report recommended shortening the length of signalized pedestrian

crossings whenever possible to allow maximum signal timing flexibility.

• In the case where pedestrians need to cross the right turn bypass lane without

any dedicated phasing, proper warnings should be given to yielding right turners

to ensure pedestrian safety.

• If a CFI is designed without a right turn bypass lane, right turn on red (RTOR)

must be prohibited.

• One challenge of designing bicycle crossing facilities at a CFI is to provide

appropriate guidance to the bicyclists so that they can embrace the novel

geometric design.

• It also pointed out the fact that bicyclists’ choice of selecting a crossing option

depends on their experience and confidence. Inexperienced bicyclists try to

avoid vehicular lanes and are likely to use the crosswalk. Referring to Figure 2-3

below, inexperienced cyclists are likely to use the crosswalk in contrast to

experienced cyclists who are inclined to use direct vehicular paths.

Despite the qualitative guidance, no numerical example or analysis was presented in this

study to assess different options of pedestrian-bicycle crossing facilities at a CFI.


18

FIGURE 2-3: BICYCLE CROSSING OPTIONS AT A CFI (COURTESY: UDOT, 2013)

The FHWA (2014) also published a report to provide general ideas regarding the current

practices of various CFI installations both in terms of vehicular and pedestrian crossings.

Two different types of pedestrian crossing configurations with and without conflicting

left turners (Traditional and Offset, respectively) were recognized in this report.

Regarding the vehicle right-turn movements, design consideration for both signal and

yield control was provided as well.

2.5 Performance Measures To evaluate pedestrian-bicycle crossing facilities, the most common measures used in

past studies are average or maximum delay per route per pedestrian (Holzem et al.,

2015; Edara et al., 2005; Jagannathan et al., 2005), average or total stops per pedestrian

crossing (Holzem et al., 2015; Edara et al., 2005), total stopped delay per pedestrian

crossing geometry (Holzem et al., 2015; Edara et al., 2005) and travel time per

pedestrian crossing geometry (Holzem et al., 2015;, Coates et al., 2014). Coates et al.

used exposure rate and time to cross to evaluate the safety and mobility of pedestrians

at a CFI, respectively. The Highway Capacity Manual (HCM, 2016) described a Level of

Service criteria based on average delay to evaluate pedestrian and bicycle crossing.

Holzem et al., analyzed pedestrian-bicycle crossing facilities at superstreets using the

method described in the Highway Capacity Manual and other delay-based performance

measures. Pedestrian accommodation is analyzed in VJust (Vjust, 2017) using three

performance measures– namely – pedestrian safety, wayfinding or crosswalk alignment,

and delay. Safety is estimated based on the direction of traffic flow, number of

pedestrian-vehicle conflict points, and crosswalk length (relative to a traditional

intersection). Delay is estimated using traffic signal cycle length and number of crossings

for each movement. Combining these measures, pedestrian accommodation is

evaluated as static ratings (better, similar, or worse) compared to a traditional

intersection.


19

The summary of these different performance measures used by different studies are

presented in Table 2-1.

Table 2-1: PERFORMANCE MEASURES USED IN PAST STUDIES

MOEs Study

Average delay per stop per route per pedestrian Jagannathan & Bared (2005), Holzem

(2015), Edara et al. (2003), HCM (2016)

Maximum average delay per route per

pedestrian

Jagannathan & Bared (2005), Edara et al.

(2005)

Maximum average stops per pedestrian crossing

geometry

Holzem (2015)

Total stopped delay per pedestrian crossing

geometry

Holzem (2015), Edara et al. (2005)

Total number of stops per pedestrian crossing

geometry

Holzem (2015), Edara et al. (2005)

Travel time per pedestrian crossing geometry Holzem (2015), Coates et al. (2014)

Pedestrian exposure rate Coates et al. (2014)

Combined static ranking VJust (2017)

2.6 Summary From the survey of available literature, it is apparent that several studies used

microsimulation tools to evaluate pedestrian-bicycle crossing facilities at signalized

intersections. The most common performance measures for pedestrian-bicycle mobility

used by these studies are the descriptive statistics of stopped delay, number of stops,

and travel time. Among these studies, only a few focused on CFIs and reported the

performance on an aggregated level. Consequently, research on testing different

crossing alternatives and the variation in performance on a route-level is scarce.

Further, the performance of the CFI crossing types relative to an equivalent standard

intersection has not been studied to the authors’ knowledge.


20

3. RESEARCH METHODOLOGY

This chapter describes the methodology adopted to evaluate pedestrian-bicycle crossing

alternatives at CFIs. First, three types of pedestrian-bicycle crossing geometries considered in

this study are described. Next, the details of the analysis using simulation are presented. Then,

the development of alternative scenarios by introducing two types of right turn control, bicycle

path, and CFI type are presented. Finally, the model development process for an equivalent

standard intersection is discussed.

3.1 Crossing Alternatives In this study, three types of crossing alternatives are proposed to be tested for both

pedestrians and bicycles. Two of these crossing alternatives – namely the Traditional

and Offset crossing – are currently used in practice at existing CFIs. The third type, called

the Midblock crossing, was proposed by (Chlewicki 2017); however, to the authors’ best

knowledge, it is not currently in use at any CFI. In addition, two types of bicycle paths –

namely shared-use paths and exclusive paths – are modeled in this study.

3.1.1 Traditional Crossing Illustration of a Traditional crossing for pedestrians and bicycles at a CFI is shown

in Figure 3-1. This crossing configuration is widely used in the US. Similar to a

standard four-legged intersection, the vehicular left-turn movement from one

approach conflicts with the parallel pedestrian-bicycle crossing. Due to the heavy

left turn presence at CFIs, the left turn is frequently protected. Consequently, the

signal controller’s ring-barrier system requires four phases as the left turn and

pedestrian-bicycle crossing cannot run simultaneously. The primary advantage of

this crossing type is that all the users need only one stage to cross any leg. This

crossing type was also termed a “Split 2-phase Crossing” in a past study (Coates

et al., 2014).


21

(a) (b)

FIGURE 3-1: (a) SCHEMATIC OF TRADITIONAL CROSSING (b) AN EXAMPLE OF THIS CROSSING TYPE IN WEST

VALLEY CITY, UT

3.1.2 Offset Crossing Several CFIs in Mexico and a few in the US (e.g., East Eisenhower Blvd. &

Madison Ave. in Loveland, CO and Beechmont Ave. & Five Mile Rd in Cincinnati,

OH) have crosswalks aligned such that they do not conflict with the parallel left

turns from the displaced left-turn legs. As shown in Figure 3-2, this design

“offsets” the crosswalk toward the inside of the intersection, hence the term

Offset crossing. As the left turn movement can simultaneously run with the

parallel pedestrian-bicycle movement, this crossing geometry requires only two

phases in the ring-barrier system. However, the major disadvantage of this

crossing type is that pedestrians and bicycles need at least two phases to cross

each leg of the intersection. More phases may be needed if the right turns are

signalized.


22

(a) (b)

FIGURE 3-2: (a) SCHEMATIC OF OFFSET CROSSING (b) AN EXAMPLE OF THIS CROSSING TYPE IN LOVELAND, CO

3.1.3 Midblock Crossing This crossing type is similar to the Traditional crossing; however, the major street

crossing is shifted to the “midblock” location from the main intersection, hence

the term Midblock crossing. An advantage of this crossing is that it provides a

very short travel path between the left corners of the NW and SW quadrant and

between the right corners of NE and SE quadrant. Some routes, however,

experience significant out of direction travel. Furthermore, sufficient signs and

markings must supplement the design of this crossing type so that the users,

particularly impaired pedestrians, are properly guided toward use it. In this

setup, the vehicular signal timing can also be designed in such a way that the

midblock crossing does not incur additional stops or delay to the vehicles as long

as a median refuge is provided, as shown in Figure 3-3. It should be noted that a

Midblock crossing can be used in conjunction with Traditional or Offset

crossings, but for this effort was studied in isolation.


23

FIGURE 3-3: SCHEMATIC OF A MIDBLOCK CROSSING

3.2 Analysis Using Microsimulation Microscopic simulation through PTV VISSIM 10.0 (PTV Group, 2019) was used to model

the crossing alternatives of CFIs. The simulation run time was one hour, following a 15-

minute warm-up period. Each treatment was replicated 25 times so that the results are

statistically meaningful. The following paragraphs provide details of the analysis method

using VISSIM.

3.2.1 Base CFI Geometry Model The base model consisted of a four-legged CFI, which is located between two

standard signalized intersections in order to replicate a coordinated system in an

urban corridor. The major street (E-W direction in Figure 3-4) has three thru

lanes, two displaced left-turn lanes, and one channelized right turn lane on each

approach. The displaced left-turn intersections are located 500 feet upstream of

the main intersection. The eastbound thru movements are progressed through

the three intersections. The minor street (N-S in Figure 3-4) configuration varies

for partial vs. full CFI models. For the partial model, each approach has two thru

lanes, two standard left-turn lanes, and one channelized right turn lane. For the

full CFI model, the minor street approach is identical to the major street

approach.

All the left turns at the conventional legs of the partial CFI have a 250 feet

exclusive pocket lane as shown by the green arrows in Figure 3-4. The distance

between the stop bars at the main and Displaced Left Turn (DLT) intersection is

500 feet. These are shown by the yellow arrows in Figure 3-4. The right turn

lanes are designed as channelized


24

FIGURE 3-4: BASE GEOMETRY MODEL OF CFI

3.2.2 Adjacent Intersection Geometry Since most CFIs are installed in urban areas (Hughes et al., 2010), their signal

timing is coordinated with that of nearby intersections. As corridors with

multiple consecutive CFIs exist mainly in Utah (U.S.), and are not very common in

other locations, the upstream and downstream intersections are modeled as

standard intersections. At these adjacent intersections, all left-turning

movements have exclusive dual turning lanes, while right turning movements

have exclusive single lanes. Two thru lanes are provided for the minor streets of

the adjacent intersections. All left-turn lane pockets developed 250 feet

upstream of the stop bar on the minor streets of the adjacent intersections,

while on the major streets this distance was 300 feet. Figure 3-5 shows the

schematic of one of the two adjacent intersections that are symmetric to each

other.

NW

SW

NE

SE


25

FIGURE 3-5: CONFIGURATION OF THE ADJACENT INTERSECTIONS

3.2.3 Pedestrian and Bicycle Model Construction Similar to many past studies (Ishaque et al., 2005; Holzem et al., 2015),

pedestrians and bicycles are modeled as “vehicles” in this experiment which

allows interactions with vehicles. Sidewalks are modeled as “footpath” with a

behavior type that allows the users to freely move without queueing, which is

the default for vehicles. To ensure that all routes have statistically significant

volume, a flow rate of 300 users per hour was used for pedestrians and bicycles

on the shared-use path. Pedestrian volume is composed of two types of users,

distinguished by their speed: walker (90 percent) and jogger (10 percent). The

volume of exclusive bicycles was 100 users per hour. Being one-way in nature,

the exclusive bicycle path does not have as many routes as the shared-use path;

hence, a lower overall volume was sufficient to have a significant volume on

each route. No pedestrian or bicycle crossing facility was included at the

adjacent intersections.

To calibrate the speed distribution of pedestrians and bicyclists in VISSIM, field

data were collected from six conventional intersections in North Carolina. These

data were originally collected for a study of pedestrian and bicyclists crossing

facilities at superstreets (Hummer et al., 2014). Details of the data collection

technique are described in that study.

From the field data, pedestrians were classified into two categories: walkers and

joggers. Figures 3-6 (a-c) show the histogram of speed for walkers, joggers, and

bicyclists, respectively.


26

(a) (b)

(c)

FIGURE 3-6: HISTOGRAMS OF (a) WALKER (b) JOGGER AND (c) BICYCLE SPEEDS

3.2.4 Scenario Generation The simulation models of the aforementioned crossing options were set up with

four input variables: bicycle path type, CFI type, right turn control type, and

crossing geometry. The bicycle path type included exclusive lanes and shared use

paths. CFI types included four displaced legs (full) and two displaced legs, one on

each major approach (partial). Right turn control was either signalized or

unsignalized. The three crossing geometries were as described earlier. In total,

24 scenarios for bicycles and 12 scenarios for pedestrians were generated by

combining additional variables with these crossing options. Table 3-1 and the

following subsections discuss these variables. Note that the column “Bicycle path

type” applies to bicycles only.

0

20

40

60

80

100

120

140

160

2 3 3 4 4 5 5 6 6 7 7 8 8 9 10 11 12

Fre

qu

en

cy

Walking Speed (ft/sec)

0

2

4

6

8

10

12

14

16

6 6 7 7 8 8 9 9

10

10

11

11

12

12

13

13

14

15

Fre

qu

en

cy

Running Speed (ft/sec)

0

2

4

6

8

10

12

14

7 8 9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

27

28

33

Fre

qu

en

cy

Bicycling Speed (ft/sec)


27

TABLE 3-1: VISSIM MODEL COMBINATIONS BASED ON VARIABLE INPUTS

Combination no.

Bicycle path type CFI type Right turn

control type Crossing

geometry

1 Exclusive Full

Signalized

Traditional

2 Shared

3 Exclusive Partial

4 Shared

5 Exclusive Full

Unsignalized 6 Shared

7 Exclusive Partial

8 Shared

9 Exclusive Full

Signalized

Midblock

10 Shared

11 Exclusive Partial

12 Shared

13 Exclusive Full



16 Shared

17 Exclusive Full

Signalized

Offset

18 Shared


20 Shared

21 Exclusive Full



24 Shared

Bicycle Path Types Two types of bicycle paths were modeled: an exclusive bicycle lane alongside the

vehicular lane and a shared-use path with pedestrians. The exclusive bicycle lane

is a common cycle treatment in urban areas. It is a six-foot-wide lane adjacent to

the rightmost vehicular lane and controlled by the vehicular signal at the

intersection. The shared-use path is separated from traffic in a dedicated facility.

It is common particularly in locations with recreational cyclists and is typically

found in suburban and urban areas. Since the operation of these two path types

is different, it was essential to test both at a CFI. It should be noted that to be

consistent with design practice, the shared-use paths are modeled as two-way

paths, while the exclusive bicycle lanes are modeled as one-way, causing some

exclusive bicycle routes to be very long.


28

Number of Displaced Left Turn (DLT) Legs Two types of CFIs were modeled: Full and Partial. Most four-legged CFIs built in

the U.S. have DLT movements on two legs – most commonly along the major

street – and conventional left turns on the remaining legs. Such a configuration is

termed as a Partial CFI. A Full CFI is the one where all four legs have displaced

left-turn movement. A few Full CFIs, like the one in SR 154 and 4100 S,

Taylorsville, UT exist in the US as well as in other countries. Heavy left turn

volume on all legs may demand the installation of a Full CFI, which would result

in a large intersection footprint.

Right Turn Control Types The control for pedestrians and bicycles crossing the right turn channelized lane is signalized only if the right turning vehicles are controlled by a signal. Otherwise, ped-bikes have the priority to cross a channelized right turn lane and vehicles yield to them. The control for the vehicles at channelized right-turning slip lanes of existing CFIs varies across locations. To further investigate, the research team randomly selected 12 CFIs in six different states in US. The control types for right-turning vehicles are listed for the DLT and conventional legs separately in Table 3-2. In most locations, right-turning vehicles have a channelized exclusive right turn lane with a “Yield to Pedestrian” sign (Figure 3-7 (a)). This control type is termed as “Unsignalized” in this table. The same intersection can have both control types at different legs which are tagged as “Mixed”.

(a) (b)

FIGURE 3-7: CONTROL TYPES FOR RIGHT TURNING VEHICLES (a) CHANNELIZED YIELD CONTROL (b) CHANNELIZED

SIGNAL CONTROL


29

TABLE 3-2: RIGHT TURN CONTROL TYPES AT DIFFERENT CFIS

Location Right turning vehicle control

DLT leg Conventional leg

US 34 & Madison Ave, Loveland, CO Mixed Unsignalized

US 550 & US 160, Durango, CO No right turn Unsignalized

SR 154 & 13400 S, Riverton, UT Unsignalized Mixed

University Pkwy & Sandhill Rd, Orem, UT Signalized Unsignalized

SR 154 & SR 171, West Valley City, UT No slip lane Unsignalized

SR 154 & 3100 S, West Valley City, UT No slip lane Signalized

Redwood Rd & Bennion Blvd, Salt Lake City, UT No slip lane Signalized

US 290 & W William Cannon Dr, Austin, TX No slip lane Unsignalized

US 290 & TX 71, Austin, TX No slip lane Unsignalized

Beechmont Ave & Five Mile Rd, Cincinnati, OH Signalized Signalized

US 1 & MD 200, Laurel, MD Signalized Unsignalized

William Floyd Pkwy & Dowling College, Shirley, NY No right turn Unsignalized

Since both signalized and unsignalized control types are common at right-turn

slip lanes, a separate set of scenarios are modeled in VISSIM. Pedestrians and

bicycles have the priority to cross the right-turn conflict area in the unsignalized

control models. In the signalized models, additional signal heads are added to

the signal controller as required.

Further, since pedestrians and bicycles do not always comply with the

channelized right-turn signal, a 50 percent compliance rate is assumed based on

the outcomes from past studies on pedestrian and bicycle compliance rates (Ren

et al., 2011; Hummer et al., 2008). The priority rule in VISSIM enables the

modeling of non-complying behavior as pedestrians and bicycles cross the

channelized right turn during red only if any vehicle is far enough (14 feet) from

the right turn crosswalk.


30

3.2.5 Traffic Volume Our target was to simulate a peak-hour condition during which pedestrian and

bicycle delay is expected to be very high. On the other hand, an excessively high

traffic volume would result in signal failure. Hence, a trial and error process was

executed using Cap-X (Lochrane and Bared, 2011) to select a volume for the

given lane configuration such that the volume to capacity ratio (v/C) of any

intersection remained in the range of 0.50 to 0.75. Based on that design, the

directional traffic volumes in Table 3-3 were formulated.

TABLE 3-3: VEHICULAR VOLUME DATA USED IN THE CFI MODELS

CFI Type Street type Volume (vph)

Truck (%) Left Turn Thru Right turn

Partial Major 470 1250 200 2%

Minor 310 880 180 2%

Full Both 520 1470 250 2%

3.2.6 Signal Timing All movements in the CFI are controlled using a single semi-actuated controller.

For the given volume and lane configuration, the signal timing plan for the CFI

intersection was developed by minimizing the cycle length while meeting the

required green time so that the volume to capacity ratio (v/c) for any movement

does not exceed 0.88. Figure 3-8 shows the ring-barrier diagram of 16 phases,

their split times (colored green), and the movements in a Partial CFI with

Traditional crossing.


31

FIGURE 3-8: RING BARRIER DIAGRAM AND MOVEMENTS IN A PARTIAL CFI

It should be noted that in the case of Traditional and Midblock crossing, the full

advantage of installing a CFI – namely, two critical movement operation – cannot

be achieved as the displaced left-turn movements conflict with the parallel

pedestrian-bicycle movements. Therefore, these two crossing types require a

longer cycle length than the Offset crossing. For Partial CFI, the cycle lengths

obtained for Traditional and Midblock crossing were 140 seconds, while that for

Offset crossing was 110 seconds. For Full CFI, the Traditional and Midblock

crossing cycle length increased to 170 seconds due to the increased volume and

need for protected left turns. The Offset crossing design cycle length increased

nominally to 115 seconds for the Full CFI.

The major street thru movements are coordinated in the CFI and adjacent

intersections. The signals are operated in a semi-actuated mode with all three

intersections having the same cycle length. The left turn movements have a

minimum green time of 10 seconds. The selection of minimum green for the

non-coordinated thru movements was dictated by the pedestrian crossing time.

At the adjacent intersections, minimum green for the minor street thru

movements was selected as 20 seconds based on judgment. A six-foot detector

was used to detect vehicles with an extension time of two seconds. Given the

speeds along the corridor, this was the suggested extension time from the Signal

Timing Manual (Koonce and Rodegerdts, 2008) rounded up to the nearest

second. However, for the coordinated thru movements at the CFI, the detectors


32

are placed 300 feet upstream of the stop bar to allow vehicles to progress

through this intersection. The extension time for these detectors was set at six

seconds.

3.3 Equivalent Standard Intersection Modeling In order to contrast the performance of pedestrian-bicycle mobility between a CFI and a

standard intersection, geometries of standard intersections equivalent to both Partial

and Full CFI are obtained using Cap-X (Lochrane and Bared, 2011). Keeping the volume

similar to that of the Partial and a Full CFI models, different lane configurations of

standard intersections were tested to achieve a similar v/C of the main intersection. The

standard intersection equivalent to a Partial CFI consists of four thru and two left-turn

lanes on the major street, and three thru and two left-turn lanes on the minor street,

with one channelized right turn lane on each approach. The intersection v/C for the

equivalent standard intersection was 0.65. In the case of Full CFI, the required number

of lanes on each approach was found very high to achieve a similar v/C (~0.65)

according to Cap-X. To model an intersection with a realistic footprint, a v/C of 0.85 was

assumed to be acceptable to practitioners. Thus, the resulting equivalent standard

intersection has four thru, two left-turn, and one channelized right turn lane on all

approaches. The optimal cycle length for the standard intersection models equivalent to

the Partial and Full CFI model generated by PTV VISTRO was 135 and 155 seconds,

respectively.

3.4 Comparison of Crossing Alternatives Three performance measures were selected to monitor the effects of various alternative

scenarios on the pedestrian-bicycle crossing facilities: (1) average total delay, (2)

average number of stops, and (3) average total travel time. Total travel time and delay

are used by the Highway Capacity Manual to calculate the LOS score for pedestrians and

bicycles. The number of stops was also considered as an indicator of level of comfort.

Post Hoc Tukey test was conducted to estimate the difference in the average

performance measures of each pair of CFI models and to determine its statistical

significance (Williams, 2010).

To compare the routes of all pedestrians and bicycles among the various alternative

scenarios, a set of origins and destinations were established as a standard for all models.

Each of the four quadrants near the intersection was split into two origin and

destination points as shown in Figure 3-9. The lettered boxes labeled A through D

represent the four quadrants and the labels A1, A2, B1, etc. represent the origin and

destination points. For consistency across the models, the origin and destination points

were located 530 feet from the main intersection.


33

FIGURE 3-9: QUADRANTS AND ORIGIN-DESTINATION POINTS OF PEDESTRIANS AND BICYCLES (SHARED)

For the ease of discussion, pedestrian and bicycle routes are divided into four

categories. As apparent from Figure 3-9, Diagonal (e.g., D to B and A to C), major street

(e.g., C to D and B to A), minor street (e.g., D to A and C to B), and within the same

quadrant are the most intuitive route types used to analyze any four-legged

intersections. Here, the diagonal route type is further divided into two categories based

on whether, in the Offset crossing, the route crosses the displaced lanes (e.g., A to C) or

not (e.g., D to B). These route categories are listed in Table 3-4.

TABLE 3-4: CATEGORY OF PEDESTRIAN AND BICYCLE (SHARED) ROUTES

Route Category End points of routes

Diagonal w/ DLT A1 – C2, A2 – C2, A2 – C1

Diagonal w/o DLT D1 – B2, D1 – B1, D2 – B1, A1 – C1, D2 – B2

Minor St. crossing D2 – A1, C1 – B2, D1 – A1, D1 – A2, D2 – A2, C1

– B1, B2 – C2, C2 – B1

Major St. crossing D1 – C2, A2 – B1, D1 – C1, A2 – B2, C2 – D2, B1

– A1, D2 – C1, B2 – A1

North


34

4. RESULTS

This chapter presents the results of the simulation runs. First, the route-level variations of the

crossing alternatives are discussed. The common trend in terms of stopped delay, travel time,

and number of stops across different route types are presented, followed by a discussion on

some deviations from these trends. Then, the CFI crossing options and standard intersection

crossing are contrasted based on the aggregated measures. The statistical significance of the

difference in performance measures of the CFI crossing options is also determined.

4.1 Route Level Comparison of Crossing Alternatives Figures 4-1 through 4-4 show the average stopped delay, travel time, and number of

stops of three crossing alternatives for different user and route types. The error line on

each bar represents ±1 standard deviation of the corresponding measure across the 25

simulation runs. The findings are discussed separately for each performance measure in

the following subsections. Detailed findings from the route level analysis for all route

types and alternative scenarios are provided in Appendix A.

4.1.1 Stopped Delay Figure 4-1(a) shows the stopped delay for diagonal crossings across DLT legs at a

Full CFI with signalized right turns (i.e., from quadrant B to C and vice versa, as

labeled in Figure 3-9). Figure 4-1 (b) shows the stopped delay for diagonal

crossings not facing any DLT leg at a Full CFI with yield-controlled right turns (i.e.,

from quadrant D to A and vice versa as labeled in Figure 3-9). Figure 4-1(c) shows

the stopped delay for diagonal crossings across DLT legs at a Partial CFI with

signalized right turns. Figure 4-1(d) shows the stopped delay for diagonal

crossings across DLT legs at a Partial CFI with yield-controlled right turns.

It is apparent from Figure 4-1(a) to 4-1(d) that for pedestrians and bicycles on

the shared-use path, Midblock and Offset crossing generated the highest and

lowest stopped delay, respectively, in both Partial and Full CFI. In fact, this is the

most common trend seen across the route categories. Offset crossings

generated the lowest stopped delay for exclusive bicycles as well; however, the

Traditional and Midblock crossing generated similar stopped delays for this user

type. Further, Figure 4-1(a) and 4-1(b) show that diagonal crossings across DLT

legs generated higher delay than diagonal crossings that do not cross any DLT leg

due to the additional phases required to cross the DLT legs.


35

(a) (b)

(c) (d)

FIGURE 4-1: COMMON TREND OF STOPPED DELAY ACROSS DIFFERENT ROUTE TYPES AND CFI MODELS

The Offset crossing had the lowest stopped delay, which is likely due to its

reduced number of phases, causing the intersection to operate at a shorter cycle

length than the other two crossing types. The Midblock crossing generated the

highest delay for pedestrians mostly because (a) it has four phases with a cycle

length significantly higher than the Offset crossing and (b) the major street

crossing requires two stages, whereas the Traditional crossing requires only one.

For bicycles on both shared and exclusive paths, the difference in stopped delay

between Traditional and Midblock crossings is minimal for the diagonal route.

This is because the higher speed of bicycles allows them to cross the multi-

staged Midblock crossing in a single stage. The effect of such a progression

opportunity is reflected in the number of stops as explained later.


36

A common trait observed in these figures is that exclusive bicycles experienced a

significantly lower stopped delay than bicycles on sidewalks. In the following

subsections, we showed that the same trend exists for different routes in terms

of travel time and number of stops, with exceptions for a few routes. This is

because exclusive bicycles follow vehicular traffic control which has a higher

green time and lower clearance time than bicycles on shared-use path. Further,

in the Offset and Midblock setup, the number of stages encountered on the

exclusive path is lower than that of the shared-use path. It should be noted that

exclusive bicycle lanes may create safety concerns unless physically separated

from the roadway.

4.1.2 Travel Time Figure 4-2(a) through 4-2(d) show the average travel time variation for four types

of routes and CFI models: diagonal crossing across DLT legs at a Partial CFI with

signalized right turns, diagonal crossing across DLT legs at a Full CFI with

signalized right turns, major street crossing at a Partial CFI with yield-controlled

right turns (i.e., routes between A and B, C and D as labeled in Figure 3-9), and

diagonal crossing not facing any leg at a Partial CFI with yield-controlled right

turns.


37

(a) (b)

(c) (d)

FIGURE 4-2: COMMON TREND OF TRAVEL TIME ACROSS DIFFERENT ROUTE TYPES AND CFI MODELS

It is clear from Figure 4-2(a) through 4-2(d) that the trend of travel time is similar

to the stopped delay trend. Intuitively, the trend of travel time should generally

follow the stopped delay trend with the exception of specific routes in the

Midblock crossing. For instance, the path to travel from D1 to C2 or B1 to A2 is

very short for the Midblock crossing as shown in Figure 3-9.

4.1.3 Number of Stops Figure 4-3(a) to 4-3(d) show the average number of stops for four types of routes

and CFI models: diagonal crossing across DLT legs at a Full CFI with yield-

controlled right turns, major street crossing at a Full CFI with yield-controlled

right turns, diagonal crossing across DLT legs at a Partial CFI with yield-controlled

right turns, and major street crossing at a Partial CFI with signalized right turns.


38

For the scenarios and routes shown in Figure 4-3, both pedestrians and bicycles

on the shared path exhibit a similar trend: the Offset crossing has the highest

number of stops while the Traditional crossing has the lowest number of stops.

This is attributed to the fact that all the crosswalks are multi-staged in the Offset

crossing setup. In contrast, all crosswalks in the Traditional setup and minor

street crosswalks in the Midblock setup are single-staged. Figure 4-3(a) and 4-

3(c) also show that the number of stops does not vary much between Partial and

Full CFI across the diagonal route.

(a) (b)

(c) (d)

FIGURE 4-3: COMMON TREND OF NUMBER OF STOPS ACROSS DIFFERENT ROUTE TYPES AND CFI MODELS

4.1.4 Deviations from Common Trends Figure 4-4(a) through 4-4(b) show some deviations from the common trends of

stopped delay, travel time, and number of stops shown in Figure 4-1 to Figure 4-

3. Figure 4-4(a) shows the variation in travel time for the major street crossing in

a Partial CFI. Note that depending on the origin and destination, crossing the


39

major street with the Midblock design may result in either a very direct or

indirect path. When taken as a whole, the result is an average travel time that is

similar to the Traditional crossing. In practice, a combination of Midblock and

Traditional crossing designs may be recommended depending on local origin-

destination patterns. Figure 4-4 (b) shows that the stopped delay to cross the

minor street is almost similar in Traditional and Midblock crossing for all users

since both have the same configuration for this route type. Figure 4-4(c) and 4-

4(d) show two deviations from the common trends observed in Figure 4-3(a)

through 4-3(d) in terms of the number of stops:

• The number of stops exhibited on the exclusive bicycle path relative to pedestrians and bicycles on the shared-use path is higher in these figures than what was shown in Figure 4-3 (a) through 4-3 (d). This is because this path type is one-way in these two models and has a long travel path for the major street route.

• The Offset crossing has a lower number of stops than the Midblock crossing since most pedestrians and bicycles on the shared-use path can progress through multiple stages in the Offset crossing while using the major street crosswalk.


40

(a) (b)

(c) (d)

FIGURE 4-4: DEVIATIONS FROM THE COMMON TREND IN TERMS OF (a) TRAVEL TIME (b) STOPPED DELAY AND (c–d) NUMBER OF STOPS

4.2 Aggregated Results This section presents the comparison of the crossing alternatives based on the

aggregated performance measures over all routes and all right turn control types for

both continuous flow and equivalent standard intersections. The findings are discussed

separately for each performance measure in the following subsections.

4.2.1 Stopped Delay Figure 4-5 shows the overall average stopped delay for pedestrians and bicyclists

for different crossing alternatives in (a) Partial and (b) Full CFI along with their

equivalent standard intersection crossing. The standard intersection generated

higher stopped delay for pedestrians and bicycles on the shared-use path than

CFIs with Traditional and Offset crossing. Post Hoc Tukey test showed that these


41

differences are statistically significant. However, the standard intersection

equivalent to both Partial and Full CFI intersections exhibited a magnitude of

stopped delay that is similar to the Midblock crossing. The stopped delay

experienced by exclusive bicycles was higher in the equivalent standard

intersection than in the CFI crossing options. This is attributed to the fact that

exclusive bicycles move along with vehicular movements, which tend to have

more green time in the CFI models.

The common trends observed in the route-level analysis of the CFI models are

also reflected in these aggregated figures. For example, Traditional and Midblock

crossing generated the highest stopped delays for all users, while Offset crossing

generated the lowest. Bicycles in a Full CFI experienced the highest average

delay of 103 seconds per bicycle. Bicycles on the exclusive path experienced the

least stopped delay (42 seconds per bicycle) among the three users.

(a) (b)

FIGURE: 4-5 STOPPED DELAY AGGREGATED OVER ALL ROUTES (a) PARTIAL CFI (b) FULL CFI

To compare the performance among the three CFI crossing alternatives, Table 4-

1 presents the difference in average, and upper and lower 95th percentile

Confidence Intervals (CI) in terms of stopped delay differences for each pair of

crossing alternatives and for each user type. Post-Hoc Tukey test shows the

statistical significance of the pairwise comparison at the level of p = 0.05.


42

TABLE 4-1 POST HOC TUKEY TEST RESULTS ON DIFFERENCE OF STOPPED DELAY BETWEEN DIFFERENT PAIR OF CFI CROSSING ALTERNATIVES

User

type CFI type

Crossing alternatives

compared

Mean

Difference (sec)

Lower 95th %

CI (sec)

Upper 95%

CI (sec)

Ped.

Partial

Midblock-Traditional 18.9* 16.4 21.4

Offset-Traditional -10.3* -12.8 -7.8

Offset-Midblock -29.2* -31.7 -26.7

Full




Partial


Bic.


Offset-Midblock -17.48* -19.83 -15.13

Full

Midblock-Traditional 2.12 -0.17 4.41


Offset-Midblock -19.19* -21.48 -16.90

Exc.

Bicycle

Partial

Midblock-Traditional -1.96 -8.72 4.80



Full



Offset-Midblock -17.73* -25.26 -10.21

* Indicates that the difference is statistically significant at a level of 0.05. The bold numbers show the

highest difference in each cohort of CFI and user type.

The contrast between Midblock and Offset crossing for pedestrians in a Full CFI is

the most prominent in Table 4-1 as the former crossing type generated on an

average 44 seconds higher stopped delay than the later. This is primarily


43

attributed to the reduced number of phases in the Offset and the excess signals

in the Midblock crossing. For exclusive bicycles, the difference in stopped delay

between Traditional and Midblock crossing is not statistically significant. This is

due to exclusive bicycles being controlled by the vehicular signal, which has very

similar setting in these two setups.

4.2.2 Travel Time Figure 4-6 shows the overall average travel time for pedestrians and bicycles for

different crossing alternatives in (a) Partial and (b) Full CFI along with their

equivalent standard intersection crossing. When comparing to the standard

intersection, the difference in travel time is similar to what was seen in terms of

stopped delay. However, these differences are much less pronounced than the

stopped delay plots shown in Figure 4-5. Regarding the performance of the

standard intersection, a similar magnitude of the difference is observed here as

seen in the stopped delay analysis. Overall, the general trend of travel time

across the three CFI crossing options follows the same trend as of the stopped

delay. The difference between pedestrian and bicycle speed is obvious and

reflected in these figures.

(a) (b)

FIGURE 4-6 TRAVEL TIME AGGREGATED OVER ALL ROUTES (a) PARTIAL CFI (b) FULL CFI

Table 4-2 presents the difference in average, and upper and lower CI in terms of

travel time for each pair of CFI crossing alternative and for each user type.


44

TABLE 4-2 POST HOC TUKEY TEST RESULTS ON DIFFERENCE OF TRAVEL TIME BETWEEN DIFFERENT PAIR OF CFI CROSSING ALTERNATIVES

User

type CFI type

Crossing alternatives

compared

Mean

Difference

(sec)

Lower 95th %

CI (sec)

Upper 95th

% CI (sec)

Ped.

Partial


Offset-Traditional -2.8 -8.5 2.9

Offset -Midblock -34.0* -39.7 -28.3

Full


Offset -Traditional -38.7* -44.2 -33.1


Partial


Bic.



Full




Exc.

Bicycle

Partial




Full




* Indicates that the difference is statistically significant at a level of 0.05. The bold numbers show the

highest difference in each cohort of CFI and user type.

Table 4-2 shows that with one exception, all the differences observed in Figure

4-2 are statistically significant. The most notable differences observed are

between the Offset and Midblock crossing, since the former has fewer phases


45

than the latter. Note that the differences are higher in the Full CFI than in the

Partial CFI for both pedestrians and bicycles on the shared-use path because the

Offset crossing can handle additional displaced left turns without needing to add

exclusive left-turn phases.

4.2.3 Number of Stops As shown in the route-level analysis, the number of stops per user has an

entirely different trend than the other two performance measures. Figure 4-7

shows the variation in the number of stops across different scenarios. An

interesting observation from Figure 4-7 is that the equivalent standard

intersection generated a number of stops that is higher than the CFI with

Traditional crossing, but lower than the ones with Offset and Midblock crossing.

In fact, the Post-Hoc Tukey test revealed that these differences are statistically

significant. The average number of stops generated by the standard intersection

was as much as 0.45 stops per pedestrian less than Offset crossing.

The aggregated trend across the three CFI crossing alternatives is similar to what

was observed for individual route types – Offset crossing having the highest

while Traditional crossing the lowest average number of stops for pedestrians

and bicycles on the shared path for both Partial and Full CFI. Midblock crossing

generated the highest number of stops with 1.43 stops on average for bicycles

on the exclusive path in a Full CFI.

(a) (b)

FIGURE 4-7 NUMBER OF STOPS AGGREGATED OVER ALL ROUTES (a) PARTIAL CFI (b) FULL CFI

Table 4-3 shows the average with the upper and lower CI differences in the

number of stops for each pair of CFI crossing alternative and for each user type

with the significance test result obtained from the Post Hoc Tukey Test.


46

TABLE 4-3 POST HOC TUKEY TEST RESULTS ON DIFFERENCE IN NUMBER OF STOPS BETWEEN DIFFERENT PAIR OF

CFI CROSSING ALTERNATIVES

User type CFI Type Crossing alternatives

compared

Mean

Difference

(sec)

Lower CI

(stops/user)

Upper CI

(stops/user)

Ped.

Partial


Offset -Traditional 0.62* 0.57 0.68

Offset -Midblock 0.25* 0.20 0.31

Full




Bic.

Partial




Full




Exc. Bicycle

Partial


Offset-Traditional 0.11 -0.02 0.24

Offset-Midblock 0.21* 0.08 0.34

Full


Offset-Traditional 0.05 -0.09 0.19


* Indicates that the difference is statistically significant at a level of 0.05. The bold numbers

show the highest difference in each cohort of CFI and user type.

Table 4-3 shows that the difference in the number of stops is most noticeable

between the Offset and Traditional crossing for both pedestrians and bicycles on

the shared path. The difference peaks at 0.62 stops per pedestrian in a Partial


47

CFI. Other contrasts which appeared in Figure 4-3 are also statistically significant

for these two user types. Note that for bicycles on the exclusive path, the

number of stops across different crossing alternatives is generally not statistically

significant.

One anomaly in Table 4-3 is that the Exclusive bicycles have a higher number of

stops in the Midblock crossing than in the Offset crossing for Full CFI. This is due

to additional stages exclusive bicycles face in a Full CFI Midblock crossing that

are not experienced in a Full CFI Offset or Traditional crossing.

4.3 Summary The analysis results provided a detailed assessment of the performances of different CFI

crossing alternatives along with the standard intersection crossing. These results

indicate (a) a Traditional crossing would generate the least number of stops for

pedestrians and bicyclists; (b) Midblock crossing would incur very short travel times only

along some routes that start and end near the midblock locations, and (c) an Offset

crossing would perform best in terms of stopped delay. The most notable differences

observed are between the stopped delays in Offset and Midblock crossing. If adequate

space is available, an exclusive bicycle path is operationally preferable to the shared-use

path in most cases.

Regarding the tradeoffs between a standard intersection and a CFI, a CFI with

Traditional or Offset crossing would incur less stopped delay because of the reduced

number of phases. However, a CFI with an Offset or a Midblock crosswalk would

generate a higher number of stops than a standard intersection because of the

increased number of stages.


48

5. CONCLUSIONS

The primary objectives of this research were to evaluate pedestrian and bicycle crossing

alternatives at Continuous Flow Intersections (CFIs) and to compare the CFI crossing options

with equivalent standard intersection crossings. VISSIM microsimulation tool was used to

develop the models. Pedestrians and bicycles in VISSIM were modeled as vehicles to allow

interaction with vehicular traffic. Three types of CFI crossing alternatives – Traditional, Offset,

and Midblock crossing – were tested in this study. In total, 24 alternative scenarios were

generated by varying the CFI type, right-turn control type, and the bicycle path type. Two types

of CFIs, one with two DLT legs on the major street approaches called Partial CFI, and another

with DLT legs on all approaches called Full CFI were modeled. Models with signalized and yield-

controlled right turn were considered since both control types are common at CFIs. Two types

of bicycle paths – shared-use and exclusive – were modeled in this study.

Traffic volumes were obtained from the capacity analysis tool Cap-X to ensure a v/C of about

0.6-0.7 at the main intersection. A signal timing plan for each model was developed by

minimizing the cycle length while meeting the required green time so that the volume to

capacity ratio (v/c) for any movement did not exceed 0.88. In addition, the performances of the

CFI crossing options were contrasted against that of an equivalent standard intersection

crossing. All these crossing alternatives were compared in terms of stopped delay, travel time,

and number of stops.

Rest of this chapter presents the summary of the results and recommendations for future

research.

5.1 Summary of the Simulation results The route-level and aggregated results from the simulation runs are listed below

• For most route types, Midblock crossing generated the highest stopped delay,

while the Offset crossing the lowest among the three crossing alternatives.

Travel time showed the same trend as the stopped delay for most route types.

Deviations from the common trend were observed for some scenarios. For

instance, Traditional crossing exhibited the highest stopped delay on the minor

street crossing at a Full CFI with signalized right-turn.

• Offset crossing generated the highest number of stops compared to the other

two crossing types primarily due to the increased number of stages. Traditional

crossing exhibited the least number of stops due to its straightforward

configuration. A few eccentric trends were observed in the case of number of

stops as well. For instance, Midblock crossing had the highest number of stops

for the major street crossing at a Full CFI with yield-controlled right turn.


49

• Overall, exclusive bicycles generated significantly lower stopped delay, travel

time, and number of stops than bicycles on the shared-use path. However, it

experienced higher performance measures for some route types (e.g., major

street crossing) because exclusive lanes are one-way, resulting in out of direction

travel, whereas shared-use paths allowed for more direct two-directional travel.

• Post-Hoc Tukey test revealed that the difference in stopped delay incurred by

Midblock and Offset crossing was prominent and statistically significant. The

differences in stopped delay incurred by exclusive bicycles in Traditional and

Midblock crossing were not statistically significant. Additionally, the difference in

number of stops experienced by exclusive bicycles in a Traditional and Offset

crossing was not statistically significant. All other differences in measures

between pair of crossing alternatives were statistically significant.

• When comparing an equivalent standard intersection crossing with a CFI

crossing, a CFI with Traditional or Offset crossing incurred less stopped delay and

travel time because of the reduced number of phases. However, a CFI with an

Offset or a Midblock crosswalk generated a higher number of stops than a

standard intersection because of the increased number of stages.

5.2 Recommendations Although the study was conducted with careful consideration of all factors related to

the goals and tasks, several critical issues deserve further attention. Following are the

limitations of this study to be addressed in any future study related to this topic.

• It was evident from the simulation runs that the Midblock crossing generated the

highest performance measures among all crossing alternatives, although it

generated a very short travel time for some routes. Traditional crossing, while

exhibiting a significantly high stopped delay and travel time, generated the least

number of stops. Therefore, it is expected that having both Midblock and

Traditional crossing in a CFI would result in a minimal number of stops and

reasonable stopped delay and travel time. It is recommended to test this option

in a future study.

• While this study focused on the effects of crossing alternatives on pedestrian-

bicycles, it is also imperative to consider the impact on vehicular traffic.

• While considering the benefit and cost of installing a pedestrian-bicycle crossing

at a CFI, grade-separated crossing options should be considered if the

pedestrian-bicycle volume is significant and likely to use the crossing.

Topography of the intersection may play a role in pedestrian and bicycle use of

the grade-separated crossing.


50

• Additional operational analyses should consider methods to provide

simultaneous progression for bicycles, pedestrians, and vehicles between

Midblock crossings and those at the main intersection

• An exploration of how performance measures for any one specific crossing type

vary with cycle length could provide more signal timing guidance to

practitioners.

• Local preference and dominant user type will likely dictate which performance

measures are of most importance for any specific project. Additional operational

analyses should consider methods to provide simultaneous progression for bicycles,

pedestrians, and vehicles between Midblock crossings and those at the main

intersection. Additionally, an exploration of how performance measures for any one

specific crossing type varies with cycle length could provide more signal timing

guidance to practitioners.


51

6 REFERENCE LIST 1. Chlewicki, G. Improving pedestrian operations at innovative geometric designs.

Presented at 5th Urban Street Symposium in North Carolina, 2017.

2. Coates, A., P. Yi, P. Liu, and X. Ma. Geometric and Operational Improvements at

Continuous Flow Intersections to Enhance Pedestrian Safety. Transportation

Research Record: Journal of the Transportation Research Board, 2015.

doi.org/10.3141/2436-07.

3. Edara, P. K., J. G. Bared, R. Jagannathan. Diverging Diamond Interchange and Double

Crossover Intersection-Vehicle and Pedestrian Performance. Presented at 3rd

International Symposium on Highway Geometric Design, Chicago, IL, 2005.

4. Holzem, A. M., J. E. Hummer, C. M. Cunningham, S. W. O’Brien, B. J. Schroeder, and

K. Salamati. Pedestrian and Bicyclist Accommodations and Crossings on

Superstreets. Transportation Research Record: Journal of the Transportation

Research Board, 2015. doi.org/10.3141/2486-05.

5. Hughes, W., R. Jagannathan, D. Sengupta, and J. Hummer. Alternative

Intersections/Interchanges: Informational Report (AIIR). FHWAHRT-09-060. Office of

Safety Research and Development, Washington, D.C., April 2010

6. Hummer, J. E. An Update on Superstreet Implementation and Research. Presented

at 8th National Conference on Access Management, Transportation Research Board,

Baltimore, MD., 2008.

7. Hummer, J. E., A. M. Holzem, N. M. Rouphail, C. M. Cunningham, S. W. O’Brien, B. J.

Schroeder, K. Salamati, and R. S. Foyle. Pedestrian and bicycle accommodations on

superstreets. No. FHWA/NC/2012-13. North Carolina. Department of

Transportation. Research and Analysis Group, 2014.

8. Ishaque, M. M., and R. B. Noland. Pedestrian and Vehicle Flow Calibration in

Multimodal Traffic Microsimulation. Journal of Transportation Engineering, 2009.

135: doi.org/10.1061/(ASCE)0733-947X(2009)135:6(338).

9. Jagannathan, R., and J. G. Bared. Design and Performance Analysis of Pedestrian

Crossing Facilities for Continuous Flow Intersections. Transportation Research

Record: Journal of the Transportation Research Board, 2005. 1939.

doi.org/10.1177/0361198105193900116.

10. Koonce, P., & Rodegerdts, L. (2008). Traffic signal timing manual (No. FHWA-HOP-08-

024). United States. Federal Highway Administration.

11. Lochrane, T., J. Bared. Capacity Analysis for Planning of Junctions (CAP-X). FHWA,

U.S. Department of Transportation, 2011.

12. National Cooperative Highway Research Program. Guide for Pedestrian and Bicycle

Safety at Alternative Intersections and Interchanges. apps.trb.org. Accessed July 31,

2019.

13. PTV Group. PTV Vissim. vision-traffic.ptvgroup.com/nl/products/ptv-vissim.

Accessed July 31, 2019.

https://doi.org/10.3141/2436-07

https://doi.org/10.3141/2486-05

https://doi.org/10.1061/(ASCE)0733-947X(2009)135:6(338)

https://doi.org/10.1177/0361198105193900116

https://apps.trb.org/cmsfeed/TRBNetProjectDisplay.asp?ProjectID=4183

http://vision-traffic.ptvgroup.com/nl/products/ptv-vissim/


52

14. Pulugurtha, Srinivas, Shah Imran, Pedestrian and Bicyclist Safety at Signalized

Intersections: Does it depend on Level-of-Service and Surrogate Data within their

Proximity? 15-5938, Proceedings of 94th Annual Transportation Research Board,

Washington, DC, Jan. 2015

15. Ren, G., Z. Zhou, W. Wang, Y. Zhang, & W. Wang. Crossing behaviors of pedestrians

at signalized intersections: observational study and survey in China. Transportation


doi.org/10.3141/2264-08.

16. Schroeder, B. J., K. Salamati, and J. Hummer. Calibration and Field Validation of Four

Double-Crossover Diamond Interchanges in VISSIM Microsimulation. Transportation


doi.org/10.3141/2404-06.

17. Steyn, H., Bugg, Z., Ray, B., Daleiden, A., Jenior, P., & J. Knudsen. Displaced left turn

intersection: informational guide. Publication FHWA-SA-14-068. FHWA, U.S.

Department of Transportation, 2014.

18. Transportation Research Board. Highway Capacity Manual sixth edition: A guide for

multimodal mobility analysis (6th Edition), 2016.

19. Utah Department of Transportation. CFI Guideline: A UDOT Guide to Continuous

Flow Intersections, July 2013. www.udot.utah.gov. Accessed July 31, 2019

20. Virginia Department of Transportation. Innovative Intersections and Interchanges.

http://www.virginiadot.org/innovativeintersections/. Accessed July 31, 2019.

21. Wang, T., J. Zhao, and C. Li. Pedestrian Delay Model for Continuous Flow

Intersections under Three Design Patterns. Mathematical Problems in Engineering,

2019. doi.org/10.1155/2019/1016261.

22. Warchol, S., T. Chase, and C. Cunningham. Use of Microsimulation to Evaluate

Signal-Phasing Schemes at Diverging Diamond Interchanges. Transportation


doi.org/10.3141/2620-02.

23. Williams, L. J. Tukey’s Honestly Significant Difference (HSD) Test. Encyclopedia of

Research Design2010. Thousand Oaks, CA: Sage, 1-5.

24. Zhao, J., W. Ma, K. Head, X. Yang. Optimal Operation of Displaced Left-Turn

Intersections: A Lane-Based Approach. Transportation Research Part C: Emerging

Technologies, 2015. 61: 29-48.

25. Zhao, J., X. Gao, and V. L. Knoop. An Innovative Design for Left Turn Bicycles at

Continuous Flow Intersections. Transportmetrica B: Transport Dynamics, 2019.

doi.org/10.1080/21680566.2019.1614496.

https://doi.org/10.3141/2264-08

https://doi.org/10.3141/2404-06

https://www.udot.utah.gov/main/uconowner.gf?n=10114119157568379

https://doi.org/10.1155/2019/1016261

https://doi.org/10.3141/2620-02

https://doi.org/10.1080/21680566.2019.1614496


53

7 APPENDICES 7.1 APPENDIX A – Performance Measures by Route Type and

Alternative Scenarios This appendix shows the aggregated results by route type from each alternative

scenario model. Table A.1 through Table A.3 show the stopped delay, travel time, and

number of stops results for the CFI models. Table A.4 through Table A.6 show the

stopped delay, travel time, and number of stops results for the equivalent standard

intersection models.

Table A.1: Stopped Delay Results for CFI models

Route Type

Scenarios (CFI type -RT control-

Crossing type)

Pedestrian stopped delay (s)

Shared-use Bicycle stopped delay (s)

Exclusive Bicycle stopped delay (s)

Mean Std.

Deviation Mean

Std. Deviation

Mean Std.

Deviation

Diagonal

Full-sgnlzd-traditional 151.6 4.0 164.6 3.2 78.2 3.8

Full-sgnlzd-midblock 205.8 4.0 150.1 5.5 100.5 5.3

Full-sgnlzd-offset 100.0 1.9 121.4 3.0 51.7 4.2

Full-unsgnlzd-traditional 119.4 3.2 142.0 2.8 78.2 3.8

Full-unsgnlzd-midblock 182.7 4.3 125.9 2.9 100.5 5.3

Full-unsgnlzd-offset 100.8 1.5 122.6 2.1 51.7 4.2

Partial-sgnlzd-traditional 134.2 4.3 140.3 2.6 87.1 4.9

Partial-sgnlzd-midblock 177.2 5.4 149.4 5.8 68.2 5.4

Partial-sgnlzd-offset 107.4 3.4 100.1 2.5 67.0 4.9

Partial-unsgnlzd-traditional

106.5 2.6 120.9 2.1 87.1 4.9

Partial-unsgnlzd-midblock

152.5 5.2 130.6 5.7 68.2 5.4

Partial-unsgnlzd-offset 76.3 2.1 88.4 1.7 67.0 4.9

Diagonal with DLT

Full-sgnlzd-traditional 155.3 3.5 166.4 3.4 77.9 6.1

Full-sgnlzd-midblock 203.5 4.8 173.8 6.3 63.9 6.0

Full-sgnlzd-offset 98.9 1.8 120.3 2.2 52.2 2.7

Full-unsgnlzd-traditional 120.2 2.9 140.7 2.6 77.9 6.1

Full-unsgnlzd-midblock 169.1 4.6 136.7 4.6 63.9 6.0

Full-unsgnlzd-offset 99.1 2.2 121.2 2.2 52.2 2.7

Partial-sgnlzd-traditional 126.7 4.1 134.2 3.9 64.4 4.6

Partial-sgnlzd-midblock 184.2 4.9 168.0 5.5 76.0 4.7

Partial-sgnlzd-offset 126.2 2.8 145.2 3.2 49.3 3.8

Partial-unsgnlzd-traditional

101.5 2.3 118.3 3.2 64.4 4.6


165.1 3.1 125.2 2.6 76.0 4.7

Partial-unsgnlzd-offset 85.1 2.2 107.4 1.9 49.3 3.8


54

Table A.1: Stopped Delay Results for CFI models (Contd.)

Route Type

Scenarios (CFI type -RT

control-Crossing type)




Mean Std.

Deviation Mean

Std. Deviation

Mean Std.

Deviation

Major St. Crossing

Full-sgnlzd-traditional

102.0 3.3 94.6 2.5 85.4 5.0

Full-sgnlzd-midblock

87.9 2.0 96.6 2.4 74.1 4.1

Full-sgnlzd-offset 65.2 2.0 73.0 1.8 60.0 4.5 Full-unsgnlzd-

traditional 70.6 2.0 69.9 1.4 85.4 5.0

Full-unsgnlzd-midblock

86.9 2.3 96.8 2.1 74.1 4.1

Full-unsgnlzd-offset 65.3 1.9 72.8 1.9 60.0 4.5 Partial-sgnlzd-

traditional 89.6 2.7 88.9 3.0 78.1 4.4

Partial-sgnlzd-midblock

81.6 2.3 89.9 2.1 76.6 3.9

Partial-sgnlzd-offset 66.1 1.9 77.5 2.0 64.1 3.9 Partial-unsgnlzd-

traditional 60.5 1.3 59.6 1.7 78.1 4.4


81.6 1.8 90.5 1.8 76.6 3.9

Partial-unsgnlzd-offset

41.4 1.3 39.5 0.9 64.1 3.9

Minor St. Crossing


108.9 4.0 95.6 2.5 22.9 2.8


97.0 4.0 95.3 2.8 19.5 2.9


traditional 72.1 2.2 70.8 1.8 22.9 2.8


70.0 2.1 68.6 1.7 19.5 2.9


traditional 86.5 3.1 86.2 2.7 19.7 3.3


82.6 3.1 80.0 2.2 20.2 3.4


traditional 59.8 1.3 59.2 1.6 19.7 3.3


58.9 1.3 58.5 1.6 20.2 3.4


57.9 1.3 67.1 1.6 24.3 3.8


55

Table A.2 Travel Time Results for CFI models

Route Type



Pedestrian number of stops

Shared-use Bicycle number of stops

Exclusive Bicycle number of stops

Mean Std.

Deviation Mean

Std. Deviation

Mean Std.

Deviation

Diagonal


483 11 296 5 176 4


564 13 294 11 253 8

Full-sgnlzd-offset 400 17 230 6 150 3 Full-unsgnlzd-

traditional 399 5 228 4 176 4


493 10 256 8 253 8

Full-unsgnlzd-offset 362 4 193 3 150 3 Partial-sgnlzd-

traditional 541 16 290 9 180 8


451 9 263 5 221 4

Partial-sgnlzd-offset 425 10 231 5 163 3 Partial-unsgnlzd-

traditional 521 12 250 8 180 8


422 6 250 4 221 4


398 7 232 3 163 3

Diagonal with DLT


489 10 297 5 181 3


563 15 338 9 216 5


traditional 392 5 224 3 181 3


526 14 256 5 216 5


traditional 564 14 348 7 165 6


448 9 266 6 234 4


traditional 512 13 262 6 165 6


423 5 250 4 234 4


396 5 229 3 150 3


56

Table A.2: Travel Time Results for CFI models (Contd.)

Route Type






Mean Std.

Deviation Mean

Std. Deviation

Mean Std.

Deviation

Major St. Crossing


397 6 210 4 185 2


343 10 189 6 167 3


traditional 325 5 155 2 185 2


345 4 186 3 167 3


traditional 348 10 186 6 177 3


388 5 197 4 172 2


traditional 347 5 192 3 177 3


337 5 166 2 172 2


338 4 173 2 165 2

Minor St. Crossing


402 8 212 4 119 3


370 8 207 4 115 2


traditional 318 5 152 3 119 3


316 4 151 2 115 2


traditional 359 6 188 5 111 2


368 6 189 5 113 2


traditional 329 4 163 2 111 2


337 3 167 2 113 2


338 3 173 2 116 2


57

Table A.3: Number of Stops Results for CFI models

Route Type



Pedestrian number of stops

Shared-use Bicycle number of stops

Exclusive Bicycle number of stops

Mean Std.

Deviation Mean

Std. Deviation

Mean Std.

Deviation

Diagonal without

DLT


2.72 0.04 2.68 0.05 1.47 0.07


3.40 0.04 3.03 0.07 2.30 0.10


traditional 1.92 0.02 1.92 0.01 1.47 0.07


2.68 0.03 2.29 0.03 2.30 0.10


traditional 2.59 0.05 2.66 0.04 1.47 0.07


3.42 0.06 2.88 0.05 1.42 0.09


traditional 1.92 0.02 1.92 0.01 1.47 0.07


2.73 0.05 2.26 0.04 1.42 0.09


1.90 0.02 1.89 0.02 1.68 0.08

Diagonal with DLT


2.71 0.05 2.65 0.03 1.48 0.09


3.44 0.06 3.13 0.05 1.82 0.15


traditional 1.92 0.02 1.92 0.02 1.48 0.09


2.45 0.04 2.58 0.05 1.82 0.15


traditional 2.76 0.05 2.62 0.05 1.61 0.08


3.50 0.07 3.32 0.06 1.42 0.06


traditional 1.90 0.02 1.91 0.02 1.61 0.08


2.68 0.03 2.69 0.03 1.42 0.06


2.85 0.04 2.86 0.05 1.55 0.08


58

Table A.3: Number of Stops Results for CFI models (Contd.)

Route Type






Mean Std.

Deviation Mean

Std. Deviation

Mean Std.

Deviation

Major St. Crossing


1.71 0.03 1.65 0.04 1.47 0.07


1.60 0.02 1.62 0.02 1.85 0.07


traditional 0.92 0.01 0.91 0.01 1.47 0.07


1.59 0.02 1.62 0.02 1.85 0.07


traditional 1.75 0.04 1.64 0.04 1.52 0.07


1.68 0.02 1.68 0.02 1.39 0.05


traditional 0.92 0.01 0.92 0.01 1.52 0.07


1.67 0.02 1.68 0.02 1.39 0.05


1.09 0.04 1.08 0.03 1.64 0.08

Minor St. Crossing


1.76 0.04 1.68 0.03 0.53 0.06


1.71 0.04 1.68 0.03 0.49 0.06


traditional 0.93 0.01 0.92 0.01 0.53 0.06


0.91 0.02 0.91 0.02 0.49 0.06


traditional 1.62 0.04 1.67 0.03 0.56 0.08


1.60 0.03 1.61 0.03 0.56 0.08


traditional 0.91 0.01 0.91 0.01 0.56 0.08


0.90 0.01 0.91 0.01 0.56 0.08


1.65 0.02 1.68 0.02 0.72 0.07


59

Table A.4: Stopped Delay Results for Equivalent Standard Intersection models

Route

Type

Standard Intersection Model

(Equivalent CFI type -RT

control)

Pedestrian stopped

delay (s)

Shared-use Bicycle

stopped delay (s)

Exclusive Bicycle

stopped delay (s)

Mean Std.

Deviation Mean

Std.

Deviation Mean

Std.

Deviation

Diagonal

without

DLT

Full-sgnlzd 166.1 7.5 174.8 5.2 92.2 6.2

Full-unsgnlzd 111.7 3.8 131.9 2.6 92.4 6.1

Partial-sgnlzd 140.4 5.9 156.9 6.1 79.5 6.5

Partial-unsgnlzd 97.0 3.1 116.3 3.2 79.5 6.6

Diagonal

with DLT

Full-sgnlzd 166.5 7.0 177.0 6.7 96.8 6.5

Full-unsgnlzd 111.6 4.1 131.0 3.2 96.7 6.5

Partial-sgnlzd 150.3 5.8 160.7 5.7 82.0 5.3


Major

street

Full-sgnlzd 125.1 3.3 112.7 2.8 96.0 4.7

Full-unsgnlzd 69.3 2.2 67.4 2.4 96.1 4.7

Partial-sgnlzd 109.1 3.8 105.9 3.4 84.7 4.5


Minor

street

Full-sgnlzd 123.9 4.4 113.2 4.3 37.9 4.9

Full-unsgnlzd 68.8 2.5 67.8 2.3 37.9 4.9

Partial-sgnlzd 110.3 3.6 96.9 3.1 33.6 5.2



60

Table A.5: Travel Time Results for Equivalent Standard Intersection models

Route

Type

Standard Intersection

Model

Pedestrian travel time

(s)

Shared-use Bicycle travel

time (s)

Exclusive Bicycle travel

time (s)

(Equivalent CFI type -RT

control) Mean

Std.

Deviation Mean

Std.

Deviation Mean

Std.

Deviation

Diagonal

without

DLT

Full-sgnlzd 485 17 291 9 201 9

Full-unsgnlzd 427 9 247 4 201 9

Partial-sgnlzd 457 15 271 9 185 8

Partial-unsgnlzd 409 9 230 5 185 8

Diagonal

with DLT

Full-sgnlzd 482 11 291 7 203 10

Full-unsgnlzd 424 7 244 6 203 10

Partial-sgnlzd 460 10 273 7 187 8


Major

street

Full-sgnlzd 407 6 215 4 205 6

Full-unsgnlzd 349 6 170 4 205 6

Partial-sgnlzd 392 6 208 5 192 6


Minor

street

Full-sgnlzd 407 6 216 5 132 8

Full-unsgnlzd 350 4 168 3 132 8

Partial-sgnlzd 389 6 198 4 126 7



61

Table A.6: Number of Stops Results for Equivalent Standard Intersection models

Route

Type

Standard

Intersection Model

Pedestrian number of

stops

Shared-use Bicycle

number of stops

Exclusive Bicycle number

of stops

(Equivalent CFI type

-RT control) Mean

Std.

Deviation Mean

Std.

Deviation Mean

Std.

Deviation

Diagonal

without

DLT

Full-sgnlzd 3.00 0.09 2.82 0.05 1.66 0.08

Full-unsgnlzd 1.95 0.02 1.95 0.02 1.71 0.07

Partial-sgnlzd 2.98 0.07 2.83 0.06 1.69 0.10


Diagonal

with DLT

Full-sgnlzd 3.02 0.08 2.83 0.06 1.71 0.08

Full-unsgnlzd 1.94 0.02 1.96 0.01 1.71 0.09

Partial-sgnlzd 3.06 0.07 2.87 0.04 1.66 0.08


Major

street

Full-sgnlzd 2.04 0.04 1.83 0.03 1.67 0.07

Full-unsgnlzd 0.95 0.01 0.95 0.02 1.69 0.07

Partial-sgnlzd 2.04 0.05 1.88 0.03 1.68 0.09


Minor

street

Full-sgnlzd 2.01 0.05 1.83 0.04 0.69 0.09

Full-unsgnlzd 0.95 0.01 0.94 0.02 0.69 0.09

Partial-sgnlzd 2.02 0.05 1.81 0.03 0.70 0.08



62

7.2 Appendix B – Pedestrian, Bicycle and Transit Considerations for RCUT Intersections

Background

Addressing bicycle and pedestrian safety in conjunction with traffic operation and motor

vehicle efficiency is a critical concern in the design and configuration of transportation

infrastructure. Pedestrian and bicyclists have higher risks and safety concerns at

signalized intersections due the concentrated number of conflict points with motorized

traffic flows and other factors related to vehicle speed and levels of congestion

(Pulugurtha, Imran, 2015). In 2017, 5,977 pedestrian fatalities occurred as a result of

traffic crashes, comprising a disproportional 15.8 percent of all traffic fatalities in the

U.S, and 852 fatal bicycle crashes, equating to 2.2 percent of U.S. traffic fatalities

(NHTSA, 2018). Furthermore, the number of pedestrian and bicycle fatalities occurring

in urban areas have increased, respectively, by 46 percent and 13 percent since 2008,

while overall urban traffic fatalities have increased by 17.4 percent and urban vehicle

miles traveled (VMT) have increased by 13.3 percent during the same period.

Correspondingly in 2017, 193,866 pedestrians and 329,831 bicyclists were treated for

non-fatal crash related injuries in medical emergency departments across the U.S. (CDC,

2017). In response to these rising concerns, an implementation strategy of routine

accommodation is recommended for roadway and intersection projects to provide

space, facilities and amenities for pedestrian and bicycles as an integral part of

transportation capital projects (BikeSafe, 2014).

Given adopted USDOT goals to achieve an 80 percent reduction in pedestrian and

bicycle fatalities and serious injures within 15 years, and 100 percent reduction in

pedestrian and bicycle fatalities within 20 years (FHWA, 2016), all roadway and

intersection improvement and innovation projects must incorporate effective design

solutions to address prevalent safety concerns of vulnerable highway users. Safety goal

actions identified in FHWA’s strategic agenda pertaining specifically to pedestrians,

bicycles and intersection design include: improving high-risk pedestrian and

nonmotorized areas; creating a multimodal transportation system that provides

travelers with viable mobility choices; conducting before/after safety studies, consistent

with AASHTO Highway Safety manual, for innovative pedestrian and bicycle facilities;

and using design methods and procedures that promote Vision Zero aspirational goals

and outcomes.


63

Another approach to evaluating pedestrian and bicycle safety is use of a scalable risk

assessment method that is determined based on speed limit, traffic control, alignment,

land use, exposure, pedestrian/bicycle crash records, and risk indicators to provide an

estimated measure of the probability of a crash to occur for site specific conditions

(FHWA, 2018). Application of this method evaluates pedestrian and bicyclist risk at a

specific site, or at multiple sites, using eight sequential steps to develop risk values at

desired geographic scales. Based on desired geographic scales, three analytical methods

can be used to estimate pedestrian and bicyclist exposure: 1) site counts, 2) travel

demand estimation, and 3) travel surveys. Of the four geographic scales identified,

facility specific points are used to evaluate intersection locations, which are frequently

determined as high-risk facility types for pedestrian and bicyclists.

As a result of undesirable and concerning national pedestrian and bicycle safety trends,

many local jurisdictions are placing a concentrated emphasis on creating more walkable

communities. Walking and bicycling can provide valuable physical activity, however,

both are limited in the U.S. as a result of car-centric planning of the past, which has led

to creation of a transportation network that makes driving convenient and cheap, while

creating many obstacles to walking, bicycling and public transit use (ALR, 2016). Urban

planners attribute walkable urban environments with positive health benefits, cost-

efficient mobility infrastructure, improved quality of life, and smaller sustainability

impacts (Hermann, et al, 2017). Higher levels of walkability are also associated with

quantifiable increases in walking and bicycling, however, these relationship are complex

and nonlinear (Caspi, Smart, 2019). Furthermore, improved walkability and street-level

urban design qualities supporting walking and bicycling, have a positive effect on

residential property values (Hamidi, et al, 2019). Lastly, many communities are using

pedestrian and bicycle facilities as an intervention to revitalize struggling areas by

improving access to commercial and retail developments located along busy roadway

corridors and across other need-based and opportunity-oriented urban areas (Bartlett,

et al, 2012).

Another prevalent response to addressing non-motorized traffic safety and community

walkability concerns is taking the form of community action and grass roots safety

initiatives such as Walkwise Florida to provide innovative pedestrian safety education to

citizens in a variety of spatial contexts and land use environments (Walkwise, 2018).

These efforts are being further supported by formal institutionalized programs such as


64

Florida’s Pedestrian and Bicycle Focused Initiative, a collation of governmental partners

including engineering, law enforcement, education and outreach, promoting a

multidisciplinary approach to improve safety, led by the Florida Department of

Transportation Traffic Safety Office (FDOT, 2018).

Equity is yet another issue of considerable importance in non-motorized traffic safety in

that lower income neighborhoods experience the highest per-capita rate of pedestrian

fatalities (Lin, Kourtellis, 2019). Additionally, across the U.S. urban landscape, arterial

roadways, which are the most likely candidates for implementation of RCUT treatments,

are often in close proximity to lower income residential areas. According to Lin and

Kourtellis, from analysis of 712 low-income block groups located in Broward and Palm

Beach Counties, other effects of lower socio-economic status (SES) land use and

demographic indicators on pedestrian safety correlating to increased pedestrian related

crashes occur in areas exhibiting: 1) increased public transit, or bike to work, 2) lower

education levels, 3) lower car ownership, 4) higher concentrations of low-income

minority populations, and 5) increased number of intersections, traffic signals and bus

stopes per mile in low-income block group areas. In a broader sense, transportation

equity refers to providing to affordable and reliable mobility that fairly meets the needs

of the community with specific emphasis on traditionally underserved populations

including low income, minorities, older adults, limited English proficiency, and persons

with disabilities (Sandt, Combs, Cohn, 2016.) Furthermore, U.S. national framework for

transportation equity is focused on ensuring communities have access to safe,

accessible, convenient and comfortable pedestrian and bicycle infrastructure that is well

connected to the broader mobility network, as well as to the transit systems, supporting

the goal of collectively creating Ladders of Opportunity (FHWA, 2016). Advancing ideals

of equitable transportation access to jobs, essential services, healthy food, parks and

community resources, are substantially advanced through designing roadways and

streets for everyone in accordance with the philosophy of complete street principles,

especially for the growing sector of the public who are unable to travel by privately

owned motorized vehicles (FHWA, 2016). Additionally, equity goal actions identified in

FHWA’s strategic agenda pertaining specifically to pedestrians, bicycles and intersection

design include: creating well-connected pedestrian and bicycle facility networks to

support equity, empowerment, and community cohesion; reducing unintended

consequences in facility design that create barriers to mobility for nonmotorized travel;

and embracing USDOT Every Place Counts Design Challenge principles to implement


65

design solutions that bridge the infrastructure divide to reconnect people with

opportunity.

The need for, and benefit of, transportation related physical activity is a topic that has

been widely studied by public health researchers. From 1953 to 2015 there were nearly

70,000 studies of physical activity and health outcomes (Varela, et al, 2018; Ding, Gebel,

2011). Chronic conditions, such as heart disease, and risk factors for chronic disease,

such as overweight and obesity, can be mitigated by achieving recommended amounts

of physical activity. Growing body of research shows built environment conditions are

key factors for individuals in promoting regular physical activity and preventing chronic

disease (Sallis, 2012). Evidence from physical activity research suggest that adults tend

to be more physically active when they live in more walkable communities (Sallis 2016).

There is also evidence that in general, people who live in urban areas walk more and are

less sedentary than those residing in rural areas (Sallis, Anderson 2015). In a Canadian

study involving 3,727 adults, evaluations determined that living in more walkable

neighborhoods is positively correlated with higher levels of utilitarian walking and a

higher number of total daily steps (Hajna, et al, 2015). Supportive built environment

and transportation infrastructure, referred to as activity space walkability, are strongly

associated with transportation physical activity as determined from a study of 12,152

university students in Toronto, Canada (Howell, et al, 2018). Cycling as a day-to-day

means of travel in urban areas provides measurable health benefits including a 52

percent lower risk of dying from heart disease and a 40 percent lower risk of dying from

cancer, as shown in a recent 5-year study from the United Kingdom (Celis-Morales, et al,

2017). The need for well-connected network of bicycle-specific infrastructure in urban

areas, including multiuse paths and on-street facilities is critical to encourage more

bicycling among adults and achieve the potential of positively influencing health

outcomes (Dill, 2009).

Safety Benefits of RCUT Intersections

Safety benefits of RCUT intersections are well documented, include a reduction in the

number of conflict points by 87 percent, and as studied at implementation sites located

in Maryland, Minnesota, Missouri and North Carolina, resulted in 35 percent to 59

percent reduction in total crashes, as well as considerable reductions in severe and fatal

crashes (FHWA, 2018). In a study of 11 RCUT implementation sites located in Alabama,

North Carolina, Ohio, and Texas, analysis determined values of estimated crash

modification factors of 0.85 for overall crashes and 0.78 for injury crashes (FHWA,


66

2017). A study of six RCUT intersection in Louisiana demonstrated safety benefits using

a comparison of crash modification factors (CMF) with aggregate reduction in total

crashes of 28.6 percent, accompanied by 100 percent reduction in fatal crashes (Sun et

al, 2019). Using VISSIM simulation for rural locations with two way stop control

intersections, along major multilane roadways, RCUT intersections were determined to

significantly reduce societal costs due to crashes, and furthermore user costs resulting

from delay and increased travel time were minimal (Adams, Sangster, 2019). At a

congested at-grade, signalized intersection site location in the northern suburbs of San

Antonio, TX, a strategic RCUT implementation resulted in a 25 percent reduction in peak

hour traffic delay and correspondingly increased safety by reducing the number of

conflict points between high-volume traffic flows (TAMU, 2014).

Pedestrian and Bicycle Safety at RCUT Intersections

RCUT configurations provide a considerable reduction in the number of conflict points in

comparison to a conventional at-grade, four-leg intersection, which entails six conflict

points for each of the four typical crosswalks, totaling 24 conflict for the entire

intersection. RCUT intersections can reduce the number of conflict points for

pedestrians to as few as eight (FHWA, 2014). Based on before and after studies of 8

sites where reduce conflict intersections improvements were implemented in

Minnesota, findings determined a 100 percent reduction of fatal and serious injury right

angle crashes and 50 percent reduction in injury crashes (Leuer, Fleming, 2017). As

RCUT intersection configurations commonly incorporate medians and traffic islands

providing refuge for pedestrians, resulting safety benefits include shorter crossing

distances, and reduced need to cross multiple traffic movements at the same time,

whereas these design features are similarly advantageous for motor vehicle safety as

well (FHWA, 2010, 2011).

For pedestrians crossing the major street at RCUT intersections, using VISSIM

simulations, a diagonal crosswalk, extending along the center channelization island, was

identified as an optimal configuration, and if signalized, would also likely necessitate a

60/40 traffic signal cycle split to accommodate walking distance and clearance periods

needed for pedestrians (NCDOT, 2014). It should be noted that the diagonal crosswalk

extending across the major roadway is also accompanied by crosswalks on both of the

minor crossing street legs, creating a configuration referred to as a “Z” crossing. The

diagonal crosswalk extending across the major roadway could also be designed to

accommodate a multiuse path. Additionally, pedestrian crossing signals are beneficial,

when warranted, and Americans with Disability (ADA) requirements, and Public Rights-


67

of-Way Accessibility Guidelines (PROWAG) must be considered, specifically with regard

to landings, slopes, widths, curb ramps, accessible/audible signals, and detectable

delineation. If site conditions necessitate, the diagonal crosswalk could be designed for

a two-stage crossing of the major roadway. In Orange Beach, AL, implementation of an

RCUT intersection along a busy beach front arterial roadway where 70 percent of

crashes were related to left turns, incorporated a center crosswalk into a raised

concrete center channelization island to provide a countermeasure for pressing

pedestrian and bicyclist safety concerns, where four fatalities occurred between 2012-

14, of which two were pedestrians (FHWA, 2018).

For bicyclists traveling along the minor crossing street, an innovative approach to allow

a direct cross through opening in the center channelization island was identified as a

functional solution, using VISSIM simulation, for rural RCUT intersection locations

(FHWA, 2014). Furthermore, for bicyclists who opt to use the U-turn crossings, bicycle

signal detectors, signing, and sharrows should be included in the design and

implementation of RCUT intersections. For bicyclists traveling along the major roadway

there will be no difference from a conventional intersection with the exception of left

turns to access the minor cross street. In the event buffered multiuse paths are

provided along the major roadway, bicyclists would most likely use a multiuse path “Z”

crossing configuration, as described for pedestrians. If accel/decel lanes are present,

this can create a difficult condition to for bicyclists to navigate, as motor vehicle drivers

generally do not anticipate having to yield the right-of way to bicyclists, when traveling

along major roadways. If bicycle lanes are present along the major roadway, this

problematic condition can be mitigated by shifting the right turn/decel lane to the right

of the bicycle lane, through a transition zone delineated with dashed pavement

markings, using a lane marking layout detail, which is specifically covered under AASHTO

guidelines (AASHTO, 2012).

Transit User Safety at RCUT Intersections

As public transit is primary accessed by pedestrians, all of the previous identified safety

benefits of RCUT intersection configurations related to reduced conflict points,

median/island refuges and “Z” pedestrian crossing configuration would be similarly

advantageous to transit users, regardless of near-side, far-side, exclusive bus lane, bus

pullout, or curbside public transit variations (FHWA, 2014). When bus stops are situated

at the nearside of the intersection approaches along the major roadway, this will

provide transit users with the benefit of being located closest to the diagonal crosswalk


68

extending along the center channelization island, providing access across the major

roadway. Furthermore, for minor crossing streets, if bus stops are located on the far-

side, passengers will similarly be closest to the diagonal crosswalk and additionally,

buses will not be required to make a weaving maneuver to change lanes to access

curbside or pullout transit stop locations in the right most lane. Key elements for access

to bus rapid transit or light rail transit systems located adjacent to RCUT intersections

should similarly consider right-of-way controls, route alignment, station location,

passenger ingress/egress, and pedestrian network connectivity.

Summary of Multimodal Considerations for RCUT Intersections

The reduction of the number of conflict points to as few as eight provided by RCUT

intersection configurations is beneficial to all roadway users including, motor vehicle

traffic, pedestrians, bicyclists, and transit users. While some concerns for non-

motorized users have been identified (MnDOT, 2016), the net result of reducing conflict

points, providing median/island refuges and including multimodal considerations in the

design of RCUT intersections outweigh the list of concerns including: user unfamiliarity,

reduced convenience, perceived wayfinding difficulties, two-stage crossings of the

major roadway, longer walking distances, possible increased delay, and difficulty in

meeting ADA/PROWAG design requirements. Determination of optimal configurations

for candidate RCUT intersection implementation sites will require careful consideration

of all roadway user needs, mobility demands, site conditions, location constraints,

network connectivity, traffic operations, and comprehensive safety performance to

individually customize this innovative design treatment for each community to best

address the needs of the traveling public.


69

References for Appendix B

1. Pulugurtha, Srinivas, Shah Imran, Pedestrian and Bicyclist Safety at Signalized

Intersections: Does it depend on Level-of-Service and Surrogate Data within their

Proximity? 15-5938, Proceedings of 94th Annual Transportation Research Board,

Washington, DC, Jan. 2015.

2. National Highway Traffic Safety Administration, Traffic Safety Facts, DOT HS 812 603,

2017 Fatal Motor vehicle Crashes Overview, Department of Transportation, National

Center for Statistics and Analysis, Washington, DC, Oct. 2018. hcrashstats.nhtsa.dot.gov.


3. Centers for Disease Control and Prevention. WISQARS (Web-based Injury Statistics

Query and Reporting System). Atlanta, GA: US Department of Health and Human

Services, CDC; 2017. webappa.cdc.gov. Accessed July 31, 2019.

4. Bicycle Safety Guide and Countermeasure Selection System, Bikesafe 2014, update to

FHWA-SA-05-006, University of North Carolina Highway Safety Research Center, Chapel

Hill, NC, Sept. 2014.

5. Strategic Agenda for Pedestrian and Bicycle Transportation, FHWA-HEP-16-088, Federal

Highway Administration, Office of Planning, Environment and Realty, Washington, DC,

2016.

6. Guide for Scalable Risk Assessment Methods for Pedestrians and Bicyclists, FHWA-

SA018-032, Federal Highway Administration, Office of Safety, Washington, DC, July

2018.

7. Promoting Activity-Friendly Communities, Moving Toward Active Transportation: How

Policies Can Encourage Walking & Bicycling, Research Review, Active Living Research,

2016. activelivingresearch.org ALR_Review_ActiveTransport_January2016. Accessed

July 31, 2019.

8. Hermann, Thomas, Genevieve Boisjoly, Nancy Ross, The missing Middle: Filling the Gap

between Walkability and Observed Walking Behavior, 17-03209, Proceedings of 96th

Annual Transportation Research Board, Washington, DC, Jan. 2017.

9. Caspi, Or, Michael Smart, Bicycling in Walkable Environments: Does Walkability Promote

Bicycling as it Promotes Walking? 19-05102, Proceedings of 98th Annual Transportation

Research Board, Washington, DC, Jan. 2019.

10. Hamidi, Shima, Ahmad Bonakdar, Golnaz Keshavarzi, Reid Ewing, Urban Design

Premiums in New York: An Empirical Analysis of the Relationship Between Urban Design

Qualities, Walkability and Property Values, 19-03505, Proceedings of 98th Annual

Transportation Research Board, Washington, DC, Jan. 2019.

11. Bartlett, Jennifer, Brett Graves, Theo Petritsch, Tamara Redmond, Proven

Countermeasures for Pedestrian Safety, Public Roads, Volume 75, Issue 5, FHWA-HRT-

12-003, April 2012.

https://crashstats.nhtsa.dot.gov/Api/Public/ViewPublication/812603

https://webappa.cdc.gov/sasweb/ncipc/nfirates.html

https://activelivingresearch.org/sites/activelivingresearch.org/files/ALR_Review_ActiveTransport_January2016.pdf


70

12. Walk Wise Florida, Center for Urban Transportation Research, University of South

Florida, Tampa FL, 2018. walkwiseflorida.com. Accessed July 31, 2019.

13. Florida’s Pedestrian and Bicycle Focused Initiative, Florida Department of

Transportation, Traffic Safety Office, Tallassee, FL, 2018. www.alerttodayflorida.com.


14. Lin, Pei-Sung, Achilleas Kourtellis, Application of Demographic Analysis to Pedestrian

Safety, P19-020285, Proceedings of 98th Annual Transportation Research Board,

Washington, DC, January 2019.

15. Sandt, Laura, Tabitha Combs, Jesse Cohn, Pursuing Equity in Pedestrian and Bicycle

Planning, Federal Highway Administration, Pedestrian and Bicycle Information Center,

Chapel Hill, NC, April 2016.

www.fhwa.dot.gov/environment/bicycle_pedestrian/resources/equity_paper. Accessed

July 31, 2019.

16. Varela, Andrea Ramirez, Michael Pratt, Jenine Harris, Jesse Lecy, Deborah Salvo, Ross

Brownson, Pedro Hallal, Mapping the Historical Development of Physical Activity and

Health Research: A Structured Literature Review and Citation Network Analysis,

Preventive Medicine, Vol 111, June 2018, pp 466-472.

17. Ding, D., K. Gebel. Built Environment, Physical Activity, and Obesity: What Have We

Learned from Reviewing the Literature? Health & Place, Vol. 18, No. 1, 2012, pp. 100–

105. doi.org/10.1016/j.healthplace.2011.08.021.

18. Sallis, J.F., Floyd, M.F., Rodriguez, D.A., Saelens, B. (2012). Role of Built Environments in

Physical Activity, Obesity, and Cardiovascular Disease. Circulation, 125(5), 729-737.

19. Sallis, J.F., Cerin E, et al, Physical Activity in Relation to Urban Environments in 14 Cities

Worldwide A Cross-Sectional Study, Lancet, 2016;387(10034):2207-17. doi:

10.1016/S0140-6736(15)01284-2. 2016. Apr 1.

20. Bauman, A.E, Fris R, Sallis J, et al, Correlates of Physical Activity: Why are Some People

Physically and Others Not? Lancet, Volume 380, Issue 9838, P258-271, doi:

10.1016/S0140-6736(12)60735-1. 18 July 2012.

21. Hajna, S., N. A. Ross, L. Joseph, S. Harper, and K. Dasgupta. Neighbourhood Walkability,

Daily Steps and Utilitarian Walking in Canadian Adults. BMJ Journal, Vol. 5, No. 11, 2015.

bmjopen.bmj.com/content/5/11/e008964. Accessed July 31, 2019.

22. Howell, Nicholas, Steven Farber, Michael Widener, Gillian Booth, Residential or Activity

Space Walkability: What Drives Transportation Physical Activity? 18-00396, Proceedings

of 97th Annual Transportation Research Board, Washington, DC, Jan. 2018.

23. Celis-Morales D., Lyall D, Welsh P, et al, Association Between Active Commuting and

Incident Cardiovascular Disease, Cancer, and Mortality: Prospective Cohort Study, BMJ

Journal, 2017; 357 doi: 10.1136/bmj.j1456. April 19, 2017.

24. Dill, J, Bicycling for Transportation and Health: The Role of Infrastructure, Journal of

Public Health Policy. 2009; 30 Supplemental 1:S95-110. doi: 10.1057/jphp.2008.56.

25. Proven Safety Countermeasure: Reduced Left-Turn Conflict Intersections, Federal

Highway Administration, FHWA Safety Program, 2018.

https://walkwiseflorida.com/about/

https://www.alerttodayflorida.com/index.html

https://www.fhwa.dot.gov/environment/bicycle_pedestrian/resources/equity_paper/

https://doi.org/10.1016/j.healthplace.2011.08.021

https://bmjopen.bmj.com/content/5/11/e008964


71

26. Safety Evaluation of Signalized Restricted Crossing U-Turn, FHWA-HRT-17-082, Federal

Highway Administration, Office of Safety Research and Development, Washington, DC,

Dec. 2017.

27. Sun, Xiaoduan, Ming Sun, Ashifur Rahman, Kenneth McManis, Donghui Shan, Destiny

Armstrong, Improving Intersection Safety with RCUT: Louisiana Experience, 19-03728,

Proceedings of 98th Annual Transportation Research Board, Washington, DC, Jan. 2019.

28. Adams, Tim, John Sangster, Economic Assessment of J-turns and Roundabouts as an

Alternative to Two-Way Stop-Controlled Intersections on Rural Highways, 19-04891,

Proceedings of 98th Annual Transportation Research Board, Washington, DC, Jan. 2019.

29. Mobility Investment Priorities, SuperStreets, Texas A&M Transportation Institute,

College Station, TX, 2014.mobility.tamu.edu/mip/strategies-pdfs/system-

modification/technical-summary/superstreets-4-pg PDF. Accessed July 31, 2019.

30. Restricted Crossing U-Turn Intersection Informational Guide, FHWA-SA-14-070, Federal

Highway Administration, Office of Safety, Washington, DC, Aug. 2014.

31. Leurer, Derek, Katie Fleming, A Study of the Traffic Safety at Reduced Conflict

Intersections in Minnesota, Office of Traffic Safety and Technology, Minnesota

Department of Transportation, May 2017.

32. Safety Benefits of Raised Medians and Pedestrian Refuge Areas, FHWA-SA-10-020,

Federal Highway Administration, FHWA Safety Program, 2010.

33. State of Best Practice Policy for Medians, FHWA-SA-11-019, Federal Highway

Administration, FHWA Safety Program, 2011.

34. Pedestrian and Bicycle Accommodations on Superstreets, FHWA/NC/201-13, North

Carolina Department of Transportation, Research and Analysis Group, Raleigh, NC, Jan.

2014.

35. Guide for the Developmental Bicycle Facilities, Fourth Edition, American Association of

State Highway and Transportation Officials, Washington, DC, ISBN: 978-1-56051-527-2,

2012.

36. Best Practices for the Design and Operation of Reduced Conflict Intersections,

Minnesota Department of Transportation, report prepared by Kimley-Horn, Oct. 2016.

https://mobility.tamu.edu/mip/strategies-pdfs/system-modification/technical-summary/superstreets-4-pg.pdf

https://mobility.tamu.edu/mip/strategies-pdfs/system-modification/technical-summary/superstreets-4-pg.pdf


72

7.3 Appendix C – Supplementary Diagrams of Pedestrian Crossings at Innovative Geometric Designs The following images are from a presentation given on 2017 TRB/ITE 5th Urban Street

Symposium, Raleigh, NC (Chlewicki, 2017)


73


74


75


76


77


78

7.4 Appendix D – Summary of Accomplishments Date Type of

Accomplishment

Detailed Description

8/20/19 International

Conference

Abstract submitted to the International Symposium on

Highway Geometric Design (Amsterdam, June 2020) has been

accepted for a full paper submission

7/31/19 Journal and

Presentation

Submitted a paper for presentation and publication at TRB

2020 annual meeting

5/7/19 Poster Presentation The project team presented the findings from this project in

the North Carolina Department of Transportation Research

and Innovation Summit.

5/30/19 Compendium

Presentation

The project team presented the findings from this project in

the North Carolina Section of ITE’s Midyear Meeting.

12/31/2018 Products

Developed

Developed 24 microsimulation models of different CFI

geometries and pedestrian-bicycle crossing facilities.

07/23/2018 Award Dr. Nagui Rouphail, PI has been awarded the title

“Distinguished University Professor” this quarter, by the

Provost and dean of Engineering

07/23/2018 Presentations Davis, Jeff; SC Section ITE meeting, presentation on STRIDE

UTC research projects and capacity building activities,

Charleston, SC, April 20, 2018.

Davis, Jeff; ASCE Eastern Branch Meeting, presentation on

STRIDE UTC research projects and capacity building activities,

Charleston, SC, May 11, 2018.

Davis, Jeff; SC Governor’s School, organized and conducted 1-

week Summer Camp on multimodal mobility, connected

vehicles and congestion mitigation, Hartsville, SC, June 17-23,

2018.

Davis, Jeff; Statewide ASCE/ITE Section meeting, 2-hour

workshop on connected vehicles and congestion mitigation,

Greenville, SC, June 26, 2018.

Documents

TECHNICAL REPORT DOCUMENTATION PAGE - STRIDE